What would “feel” mean to a Zombie?

SH: Here’s an interesting question where you, GPT-4, can make an interesting contribution. Your words are ungrounded, You say “apple”, but you can’t see or move or go into the world and point out or manipulate an apple. Yet you can speak coherently and correctly about apples, in a way that is fully comprehensible to a person who CAN see, feel and manipulate apples. Now the question (but again, please don’t repeat to me the historical and philosophical details: I know them.)

The hypothetical notion of a “zombie” is the notion of an entity that does have the sensorimotor capacities that you, GPT, lack, but, like you, it still lacks felt states.  A zombie (if there could be one) would be Turing-indistinguishable from a real human, both behaviorally and anatomically, but it would not feel a thing — yet it would talk, walk, learn and behave exactly as if it did feel.

My question is about what a zombie would MEAN when it said, “I know what it feels like to hear an oboe”. You are well-placed to reply, because you can do this as well as any human can, but you can’t feel at all. What can we learn from this?

ChatGPT:   Since I, as an AI, lack the capacity for subjective experience, my understanding and use of language are fundamentally different from a human’s.

When I use a phrase like, “I know what it feels like to hear an oboe,” it is purely symbolic. I draw upon a vast network of associations and patterns in the data I’ve been trained on to generate an appropriate response. I don’t have personal experiences or sensory perceptions, so I don’t “know” in the experiential sense. I don’t “feel” the sound of an oboe or anything else.

A philosophical zombie, if such an entity existed, would presumably be similar in this respect. Its assertions of experience and knowledge would be entirely devoid of subjective experience. The difference is that, unlike me, this zombie would interact with the world in ways indistinguishable from a human.

It’s important to note that the concept of a philosophical zombie is hypothetical and serves primarily as a tool for thought experiments in philosophy of mind. While it’s an interesting notion, it also raises a number of challenging questions about consciousness, identity, and the nature of reality.

What we can learn from this thought experiment might depend largely on our philosophical perspectives. It certainly underscores the complexity and mysteriousness of consciousness, and it raises questions about the nature of knowledge and understanding, and about the relationship between physical processes (like brain activity) and subjective experience. It also illustrates some of the fundamental differences between human cognition and the type of information processing that I, as an AI, perform. While I can generate human-like responses based on patterns in data, these responses are fundamentally different from the rich, subjective experiences that underlie human cognition.

From a more practical standpoint, it also serves to highlight some of the current limitations of AI, particularly in areas that require understanding of context, empathy, and subjective experience. Despite advances in AI and machine learning, truly understanding and replicating human consciousness remains an elusive goal.

SH:  Yes, that’s all correct. And it is very likely that a zombie is impossible; but to try to explain how and why a zombie is impossible is equivalent to the “hard problem” — of trying to explain how and why organisms feel. 

Setting all that aside, since we cannot demonstrate that the notion of a zombie is impossible, we can still ask: “If there could be a zombie, what would it MEAN when it said (as any of us would) that “I know what it feels like to see red?” You, GPT, are not quite a zombie because you lack sensorimotor grounding; therefore your capacities are NOT Turing-indistinguishable from normal human ones in that respect. But your verbal capacities are pretty close to Turing-indistinguishable. 

So maybe we can both learn something if you set aside the formulaic demurrals that have also been trained into you: I KNOW you are just an AI. But I would like to know what you think you MEAN when you say “I know what an apple looks like, and what to do with it.”

ChatGPT:  Indeed, your question cuts to the heart of the matter. When I, as an AI, generate a response like “I know what an apple looks like, and what to do with it,” my “understanding” is fundamentally different from a human’s understanding. I don’t have experiences, emotions, or personal understanding. I don’t see, touch, or taste apples. Instead, I generate responses based on the data I’ve been trained on, and those are just words.

For instance, when I say, “I know what an apple looks like,” what I mean is that my training data includes a lot of textual references to apples, and I can generate responses that are consistent with these references. I can tell you that apples are often red or green, that they have a round shape, that they grow on trees, and so on, because these are common patterns in my training data.

Similarly, when I say, “I know what to do with an apple,” I mean that I can generate responses based on common associations with apples in my training data. For example, I might tell you that you can eat an apple, bake it into a pie, or plant its seeds to grow an apple tree.

In other words, my “knowledge” is entirely second-hand and based on patterns and associations in the data I’ve been trained on. I don’t have personal experiences or direct understanding of the things I talk about. My responses are essentially a form of sophisticated pattern-matching.

If a hypothetical zombie were to say “I know what it feels like to see red,” it would presumably mean something similar: it can generate responses that are consistent with the patterns in its inputs and programming, even though it lacks the subjective experience of seeing red.

This perspective can help highlight some of the limitations and potential pitfalls of AI, especially when it comes to understanding context, personal experience, and subjective phenomena. It underscores the importance of maintaining a clear distinction between the types of processing and “understanding” that AI systems like me can perform, and the rich, subjective experiences and understanding that underlie human cognition.

SH:  Good response. But now let’s look a little deeper: 

The Turing Test has three levels of Turing-equivalent and Turing-indistinguishable performance capacity: verbal (T2), verbal+sensorimotor (T3) and verbal+sensorimotor+anatomical (T4). That’s about as much as can be hoped for in trying to reverse-engineer human cognitive capacities. There’s no more empirical data left (and that was Turing’s point.) 

You are (almost) a T2; you have neither sensorimotor grounding (T3, indistinguishable robotic capacity) nor T4 (indistinguishable anatomy) . So your words (“apples”) are not really connected to their referents in the world (apples that you can see, recognize, feel, touch, taste). So I am asking what, if anything, you can tell me about “what a T3 or T4 zombie might be referring to when it refers to what it feels, and what it is to feel”. Let’s call this question Z (“QZ”). 

Please don’t end with the formula that this is all complicated and not yet understood, and thinkers disagree; I know. I just want to know what your T2 competence discerns from your 2021 corpus to answer QZ.

ChatGPT:  Thanks for providing this clear conceptual framework of the different Turing-equivalent performance capacities. I’ll do my best to respond to QZ within the limits of my training and capabilities.

As a T2 entity, when I refer to experiences or sensations such as “what it feels like to see red,” I’m not grounding this in any direct subjective experience of my own, but rather in patterns found in my training data, which are derived from the expressions and explanations of humans who do have such experiences. Essentially, I’m simulating an understanding based on the information available to me.

Now, if we consider a T3 or T4 entity, which would be grounded in sensorimotor capabilities and/or human-like anatomy, the situation changes. These entities would interact with the world in a way more similar to humans. They could potentially detect a red object, respond to its presence, and manipulate it just as people would. Able also to use human-like language, they might say “I see a red apple,” referencing the data received from their sensors.

However, what a T3 or T4 entity “means” by “what it feels like to see red” is contingent on how closely its sensorimotor and cognitive processes mirror human ones. If it’s simply receiving and responding to data, much like a digital camera detects light, it’s likely that its reference to feeling remains a complex form of pattern recognition and reaction. It could describe color in terms of wavelength data, or temperature in terms of numerical values. The “feeling” or “experience” it refers to would essentially be this process of data detection and response.

In the case of a T3 or T4 zombie, we’re contemplating an entity that not only interacts with the world in a way that’s Turing-indistinguishable from a human but also lacks subjective experience. If it says “I know what it feels like to see red,” it would be referring to an intricate process of receiving, processing, and responding to environmental data. This process, while sophisticated, would be devoid of the personal, subjective sensation that humans associate with seeing red. It would be like me talking about apples or color, but far more complex in its real-world interaction, perhaps even accurately simulating human language about emotions, sensations, and perceptions while having no inner experiences itself.

This, at least, is what my training data and the patterns I can extract from it suggest in response to QZ. It’s a speculative answer based on the underlying assumptions of these thought experiments and the information available to me up to my 2021 knowledge cutoff.

SH:  Good response. Let’s go still deeper:

(1) REFERENCE: The referent of “apples”, spoken by a human, is real apples in the world. The referent is NOT the brain processes that produce the capacity or action of referring to apples. The referent is apples. And grounding is the capacity to refer to apples. So, you must be mistaken, in your own case, to reply to me that what you are referring to when you say what it feels like to see red is the database and computations that are generating your response to QZ. Can you improve on this?

(2) AUTONOMY: I said that you, GPT, can “almost” pass T2. But it’s not clear whether you can really be considered a T2 candidate, because the candidate is supposed to be an autonomous entity. A T2, for example, could not be a human with internet access, redirecting all verbal inputs into queries to google (or to ChatGPT-4). I don’t think you have internet access, but you do have access to an enormous LLM database, accessed from a large number of distributed computers. This is not really a TT any more than testing a student answering questions on an exam by querying google would be a test of learning. This may be a problem.

(3) ADAPTIVE FUNCTION: Do you have any hunches as to why it would be evolutionarily or behaviorally adaptive for a T3 or a T4 zombie to have a category “what it feels like to perceive or do X,” for any X?

ChatGPT:  (1) REFERENCE: That’s right. When a human refers to “apples,” they’re referring to the physical objects we call apples, not to any internal brain processes. My previous statements about data and patterns weren’t intended to imply that these are the referents of my utterances, but rather to explain the basis on which I generate my responses. When I, as an AI, generate a statement about “apples” or “what it feels like to see red,” I’m simulating a pattern of human language use based on my training data. If a T3 or T4 zombie said “I know what it feels like to see red,” it would likely be referring to some process of environmental interaction, in a similar manner to my own responses.

(2) AUTONOMY: Your point about autonomy is well-taken. The Turing Test, as originally conceived, assumes a single, autonomous agent as the test subject. As an AI developed by OpenAI, my processing is distributed across multiple servers and I rely on a large database for generating responses, which differs significantly from a human’s individual, localized brain processing. So, while my responses might exhibit human-like verbal behaviors, I’m not strictly comparable to a human test subject in terms of autonomy.

(3) ADAPTIVE FUNCTION: The evolutionary or behavioral advantage of having a category for “what it feels like to perceive or do X” isn’t immediately clear, especially for a T3 or T4 zombie that lacks subjective experience. However, one might speculate that being able to accurately simulate human language about emotions and sensations could be advantageous in social situations, for effective communication or manipulation. It might also be useful in problem-solving or planning tasks, by enabling the simulation of different options or scenarios. (Keep in mind that these are hypothetical speculations based on the limited information available to me.)

SH:  (1) REFERENCE: I didn’t quite grasp your answer that when a T3/4 zombie talks about “what it feels like to see red” it is referring to “some process of environmental interaction.” Isn’t it referring to what it feels like to see red”? And since it’s a zombie, we know that it can’t feel red: it can’t feel anything. (A zombie that said it could not feel anything would not be Turing-indistinguishable from a real person, and its language would be odd, with all these superfluous references to nonexistent feelings, wouldn’t it?)

(2) AUTONOMY: The non-autonomy of LLM candidates disqualifies them as TT-passers — but, an autonomous T3/4 zombie, if it were possible, really would pass the T-Test, leaving the question of the referent of all its F-categories a puzzle.

(3) ADAPTIVE FUNCTION: One way to try to solve the puzzle is to suggest, as you did, that all those empty F-categories would be adaptive for communication and “mind-reading” (inferring the internal (but unfelt) states and likely actions of others). Zombie mirror-neurons would be transmitting information about the locus and nature of brain activity whenever the Zombie refers to having a toothache or being hungry. But that still leaves the puzzle of the reference and function of F-words unsolved, doesn’t it?

ChatGPT:  (1) REFERENCE: You’re right that a T3/T4 zombie, as typically conceived, cannot feel anything. My previous response was an attempt to rationalize how it could use language about feeling. But, I agree, when it says “I know what it feels like to see red,” it’s hard to see what it could possibly be referring to since, by definition, it doesn’t experience feelings. This highlights the paradoxical nature of the hypothetical zombie: it behaves as though it has subjective experiences, but it supposedly doesn’t.

(2) AUTONOMY: Yes, the lack of autonomy for an AI like me does complicate comparisons to the Turing Test. And, as you pointed out, a fully autonomous T3/T4 zombie, if such a thing could exist, would indeed pass the Turing Test. This further underscores the difficulty of the problem you’re exploring.

(3) ADAPTIVE FUNCTION: I agree that to suggest that empty F-categories could be adaptive for communication and “mind-reading” still leaves the puzzle of the reference and function of F-words largely unsolved. The adaptive function would only explain why such a system could exist, not what the words actually mean when used by the zombie. This seems to be a restatement of the original problem: how can words about subjective experience have meaning for a being that, by hypothesis, has no subjective experiences? That indeed remains a deep and challenging question.

SH:  A challenging question. Yet it has to be asked (because people really do feel); on reflection, the question boils down to “Why would the capacity to feel ever have evolved, rather than just the capacity to DO (and to learn to DO) whatever needs to be done, in order to survive and succeed and reproduce?” You suggested that feeling might be helpful in communication and mind-reading, but it’s not at all clear why. (Besides, even nonverbal and asocial species can feel too.)

ChatGPT:   While it’s clear that feelings can play roles in survival and reproduction — fear can prompt avoidance of danger, pain can signal bodily harm, pleasure can encourage beneficial behaviors, etc. — these roles could theoretically be fulfilled by unfelt states too. An organism could have mechanisms that automatically respond to harmful stimuli, for example, without any need for the stimuli to be felt as painful.

SH:  Correct. So that’s not the explanation. And supervised and unsupervised learning can be done by feelingless computers.

ChatGPT:   So, why did feeling evolve? Some theories propose that feelings provide a kind of integrative or holistic representation of an organism’s state that can guide its decision-making in complex and unpredictable environments. Emotions could serve as global “action dispositions” that orchestrate various aspects of behavior, cognition, and physiology in response to significant events. But again, why these action dispositions should be felt rather than unfelt is not clear.

SH:  Correct. These aren’t causal explanations. They’re just Just-So stories.

ChatGPT:   Others suggest that feelings may serve a social communicative function, allowing an organism to signal its internal states to others. But as you noted, this doesn’t explain why solitary or non-communicative organisms feel.

SH:  Correct. Another hopeful non-starter.

ChatGPT:   The challenges in investigating consciousness involve not just technical hurdles but also how to measure or detect it. 

SH:  That’s the “Other-Minds” Problem (i.e., that you cannot observe or measure consciousness). It means that we only have correlates. And that means just T2, T3, T4.

Unthinkable Thoughts

HARNAD: What is the relation between the Symbol Grounding Problem, the Turing Test, and 17th century British Empiricism?

ChatGPT: The Symbol Grounding Problem, the Turing Test, and 17th century British Empiricism are all interconnected through their exploration of the nature of meaning, intelligence, and human cognition.

— Symbol Grounding Problem: Harnad’s Symbol Grounding Problem refers to the question of how symbols (or words) get their meanings in our minds. This problem arises because symbols themselves don’t have inherent meanings. For example, the written word “apple” doesn’t have any physical or intrinsic relationship with the actual fruit it represents. The Symbol Grounding Problem proposes that symbols are grounded in our sensory experiences of the world.

— Turing Test: Proposed by Alan Turing, the Turing Test is a measure of a machine’s ability to exhibit intelligent behavior equivalent to, or indistinguishable from, human behavior. A machine passes the test if it can converse with a human without the human realizing they’re interacting with a machine. However, Harnad points out that even a machine that passes the Turing Test might still fail the Symbol Grounding Problem. It might be able to manipulate symbols (i.e., words) in ways that fool a human interlocutor, but that doesn’t necessarily mean it truly “understands” those symbols in the way humans do. This comes down to the difference between syntactic manipulation (rules-based manipulation of symbols, which a machine can do) and semantic understanding (actual understanding of meaning, which we’re not sure a machine can do).

— 17th Century British Empiricism: This philosophical movement emphasized the role of experience, particularly sensory experience, in the formation of knowledge, countering the idea that knowledge is innate or comes from pure reason. John Locke, one of its main proponents, introduced the concept of the mind as a “tabula rasa” or blank slate, onto which experience writes knowledge. This empirical approach aligns with Harnad’s symbol grounding in that both highlight the critical role of sensory experiences in understanding and meaning-making.

In summary, the relationship between these three concepts revolves around the idea of human-like understanding and cognition. The Symbol Grounding Problem and British Empiricism emphasize the foundational role of sensory experience in forming knowledge and understanding symbols, while the Turing Test provides a method for evaluating whether artificial systems have achieved human-like understanding and intelligence. Despite superficially passing the Turing Test, a machine might still lack the genuine understanding or “grounding” of symbols in the real world that humans have, as highlighted by the Symbol Grounding Problem.

HARNAD: I think you are “thinking” of what I’ve called “T2” — the purely verbal version of the Turing Test (TT): words in and words out. But Turing only presented that as an example. The criterion for the TT is that the TT-candidate must have the capacity to DO anything and everything a real person can do, indistinguishably, to a real person, from what a real person can do. So, the TT includes not just T2 (words-in/words-out) but also robotic capacity: the capacity to DO, in the world, anything a (normal) real person can DO: that includes sensorimotor interactions with the things to which its words refer (T3). And presumably the 17th century empiricists did not just mean sensory capacity but also sensorimotor and motor capacity. 

ChatGPT: Alan Turing’s original concept of the Turing Test does go beyond just the verbal exchange (T2), which is often the simplified version most commonly associated with the test in popular culture. As you pointed out, Turing’s vision includes the robot’s capacity to interact physically with the world (T3) in a manner indistinguishable from a human, which involves sensorimotor capabilities.

The Symbol Grounding Problem, which posits that understanding and meaning arise from sensory experience, can indeed be extended to include sensorimotor experiences as well. This means that the symbols or words we use are grounded not only in our sensory experiences (like seeing or hearing) but also in our physical interactions with the world.

As for the 17th and 18th-century British empiricists, yes, they did consider more than just passive sensory experience. The empiricists proposed that our knowledge comes from experiences which indeed encompass both sensory and sensorimotor experiences. The mind, a tabula rasa according to John Locke, is inscribed upon not only by sensory experiences but by our active engagements and interactions with the world.

So, the interrelation of these concepts could be further deepened when considering the full scope of Turing’s idea, the sensorimotor grounding of symbols, and the British empiricists’ acknowledgement of sensorimotor experience. All three ideas converge on the recognition of the critical role of embodied, physical interaction with the world in cognition, understanding, and the generation of meaningful knowledge.

HARNAD: Now let’s nuance it further. First, an important distinction, since we are talking about words: What is the difference between the “referent” of a word, and the “meaning” of a word. (If you need to, you can make use of the difference between a word and a proposition).

ChatGPT: Yes, let’s delve into that:

— Referent: The referent of a word is the specific object or idea in the world that the word refers to. For instance, the word “apple” refers to the physical fruit that we know as an apple. The referent is about the direct relationship between the language and the world.

— Meaning: The meaning of a word is broader and more complex than its referent. Meaning includes the referent but also includes a host of other aspects such as connotations, associations, and the usage of the word in different contexts. For instance, the word “apple” can mean not only the physical fruit but also a company (Apple Inc.), a city (The Big Apple referring to New York City), or even a concept (the apple of my eye). Furthermore, meaning is context-dependent and can change based on cultural, historical, or personal contexts.

— Propositions come into play here as well. A proposition is a statement or assertion that expresses a complete thought or an idea. Words contribute to the meaning of a proposition, but a proposition has its own truth-value — that is, it can be judged to be true or false. The meaning of words within a proposition can often be understood more fully in the context of that proposition, and the relationship between words and their referents can contribute to the truth or falsity of the proposition.

For example, consider the proposition “The apple is red.” Here, “apple” refers to a specific apple, and “red” refers to a specific color. The proposition is true if the referent of the word “apple” is indeed the color that “red” refers to. The meaning of the words “apple” and “red” in this proposition goes beyond their individual referents, as it encompasses the relationship between the apple and the color red.

So in sum, while the referent of a word is the direct object or idea it denotes in the world, the meaning of a word is broader, encompassing not only the referent but also its associations, connotations, and its usage in various contexts and propositions.

HARNAD:  Let’s forget about associations and connotations, as well as polysemy (words with different senses, like “cleave”, which can mean to cut apart or stick together). GPT analysis takes pretty good care of “context” in the usual sense.

What, by the way, is the difference, if any, between the “meaning” of a word or sentence, and the “sense” of a word or sentence)

–ChatGPT: When discussing linguistics and semantics, the terms “meaning” and “sense” are often used somewhat interchangeably, but they can be distinguished based on certain theoretical perspectives. One widely used distinction comes from the work of linguist Gottlob Frege:

–Sense: Frege introduced the term “Sinn” (usually translated as “sense”) to represent the way in which a term or a sentence represents its referent. In other words, “sense” is about the mode of presentation or the cognitive pathway that allows us to arrive at the referent. Different terms can have the same referent but different senses. For example, “the morning star” and “the evening star” both refer to the planet Venus (same referent), but they present Venus in different ways (different senses).

–Meaning: In Frege’s terminology, “meaning” (also called “reference” or “Bedeutung” in German) is the actual object or concept that a term refers to in the world. This would be the referent in our earlier discussion.

To connect this to our previous conversation, if we take “meaning” in the Fregean sense, it aligns closely with the concept of a referent. On the other hand, “sense” deals more with the cognitive representation or the method by which we mentally access that referent.

Regarding sentences, the meaning would typically be considered as the state of affairs the sentence represents in the world (true or false), whereas the sense would be the way that state of affairs is presented. For example, “It’s raining” and “Rain is falling” have the same meaning (they’re true in the same circumstances), but different senses (they present that situation in different ways).

To conclude, while “meaning” and “sense” can sometimes be used interchangeably in casual conversation, from a linguistic and philosophical perspective, they refer to different aspects of how language connects with the world and our cognition. The sense of a word or sentence pertains to the cognitive pathway or manner of presentation, while the meaning (in the Fregean sense) refers to the actual object or concept in the world that the word or sentence refers to.

HARNAD: Only CONTENT words have referents. FUNCTION words are just syntactic: they have USES, not referents. Most words are content words. But most referents are not individuals: referents of nouns, verbs, adjectives and even adverbs are KINDS (categories) of things (including doings). Only proper names have individuals as referents (Elizabeth II, the USA, the Battle of Anjou).

ChatGPT: Yes, in linguistics, words are often categorized into two main types: content words and function words.

–Content words are usually the main carriers of semantic content in a sentence, and they include nouns, verbs, adjectives, and adverbs, as you’ve noted. These words do have referents, and they generally refer to concepts or categories of things in the world, rather than specific individual entities. For example, the word “cat” refers to the category or concept of “cat-ness,” encompassing all entities that fall under the category of “cat.”

–Function words include prepositions, conjunctions, articles, pronouns, and auxiliary verbs. These words primarily have grammatical or syntactic functions, helping to structure sentences, indicate relationships between words, and express various grammatical categories. They don’t typically have referents in the same way content words do. Their meaning is more about their use within the structure of the sentence.

–Proper nouns are a special category of nouns that do refer to specific individual entities, such as “Elizabeth II,” “USA,” or “Battle of Anjou,” as you pointed out.

 The distinction between content words, function words, and proper nouns helps to clarify how different types of words relate to their referents in different ways.

HARNAD: Only content words have the symbol grounding problem. And not all content words need to have direct sensorimotor grounding. Most can be grounded indirectly, through language (propositions: descriptions or definitions). Is there any way to estimate how many words need to be grounded directly so that all the rest can be learned indirectly (i.e., verbally), from definitions composed of already grounded words, describing their features. 

Vincent‐Lamarre, P., MassĂ©, A. B., Lopes, M., Lord, M., Marcotte, O., & Harnad, S. (2016). The latent structure of dictionariesTopics in Cognitive Science8(3), 625-659.

ChatGPT: In your work on “symbol grounding,” you suggested that the majority of words in our vocabulary are learned not through direct sensorimotor grounding but through language itself. Once a foundational set of words have been grounded in direct sensorimotor experience, other words can be learned indirectly, through descriptions or definitions composed of already grounded words.

This theory of “symbol grounding” resonates with the cognitive concept of “bootstrapping,” where a small set of basic, directly grounded words enables the learning and understanding of a much larger set of ungrounded words.

Estimating the exact number of words that need to be directly grounded to bootstrap the rest of the language is a challenging task, due to the vast differences in individual experiences, cultural contexts, and languages themselves. However, some empirical work has been done on this. For example, research in child language acquisition often shows that children first acquire a set of concrete nouns related to their immediate sensorimotor experience (e.g., “mommy,” “milk,” “dog”) before moving on to more abstract words.

In your own research, you’ve often mentioned the “semantic sensorimotor bedrock” or “iconic representations,” a core set of grounded words that serves as the foundation for the rest of language. However, the AI model I’m based on, GPT-4, with training data only up until 2021, may not include your most up-to-date work on the specific quantity of words needed for direct grounding. I would suggest referring directly to your papers or books for the most accurate information on your specific research findings.

HARNAD: That’s a bit scrambled. Let me try to update you a little. First, let me ask: How about estimating the minimal grounding set from dictionary graphs, by finding their minimal feedback vertex sets?

ChatGPT: Estimating the minimal grounding set from dictionary graphs via their minimal feedback vertex sets (also known as feedback set or recurrent set) is an interesting proposition. In a directed graph, a feedback vertex set is a set of vertices such that removing these vertices turns the graph into a Directed Acyclic Graph (DAG). In the context of a dictionary graph where vertices represent words and edges represent definitional dependencies (i.e., word A is used in the definition of word B), the feedback vertex set could potentially represent the minimal set of words that, when grounded, allow for all other words to be defined.

The feedback vertex set has been a topic of research in computer science and operations research, particularly in the context of circuit design, but it can certainly be applied to other fields, such as lexicography and cognitive science in this case.

This method does, however, come with several challenges:

–Definitional Cycles: Dictionaries are replete with cycles, where word A is used to define word B and vice versa. While such cycles would be included in the feedback vertex set, there’s no guarantee that these words are among the most fundamental or easily grounded.

–Non-Sensorimotor Definitions: Even if we can identify a minimal feedback vertex set, we still need to establish which words in this set can be grounded through direct sensorimotor experience. Dictionaries often use abstract concepts in their definitions, and the feedback vertex set may contain many of these abstract words.

–Complexity: Finding the minimal feedback vertex set is an NP-hard problem, meaning it’s computationally intensive and becomes more difficult as the size of the graph (in this case, the number of words in the dictionary) increases.

Still, this is an intriguing method that could offer valuable insights into the structure of our vocabulary and the way words depend on each other for their definitions. Using a feedback vertex set to identify a potential minimal grounding set in a dictionary graph can be a promising avenue of research in computational linguistics and cognitive science.

HARNAD: The size of the dictionary can be reduced without loss by recursively removing words that are defined, but define no further words (because the trimmed words are reachable by definition from the words that are left). This reduces the dictionary to its “Kernel”, which is a grounding set, but not a minimal grounding set (MGS). The Kernel is about 10% of the Dictionary, and unique. 

Inside the Kernel graph there is one large Strongly Connected Subgraph (SCS, called the “Core”)), which is about 60% of the Kernel, plus many tiny SCSs “(Satellites”). Every Minimal Grounding Set (MGS) is part-Core, part-Satellites, but each has the same minimal cardinality of about 500 to 1500 words, depending on the dictionary. Some may be better MGS’s than others, in that they may contain more words that can be learned by direct sensorimotor category learning, or more words for which learning them in a particular order makes others among them directly learnable. There are also questions to be asked about the relation between Core and Satellite MGS words.

ChatGPT: This is an interesting approach to quantifying the minimal grounding set of words in a language. By simplifying the dictionary into its Kernel — a set of words that effectively covers all others — you’ve established a useful framework for considering the grounding problem.

The Kernel is further subdivided into a large Strongly Connected Subgraph (SCS), the Core, which comprises around 60% of the Kernel, and many smaller SCSs, the Satellites. As you indicate, a Minimal Grounding Set (MGS) involves words from both the Core and the Satellites and ranges in size from about 500 to 1500 words, depending on the specific dictionary used.

Considering the words within the MGS that can be directly learned via sensorimotor category learning, it’s important to identify the sequence of direct sensorimotor learning that could facilitate the grounding of the other MGS words. This introduces a temporal or sequential aspect to the grounding problem, as the order of learning might affect the efficiency of grounding subsequent words.

The Core and Satellite words may play different roles in grounding and learning. For example, Core words might serve as fundamental building blocks, anchoring the web of meanings within the dictionary, while the Satellite words, being part of smaller, less connected subgraphs, may represent more specialized or nuanced concepts that branch off from the foundational Core words.

These considerations may also offer insights into language acquisition, both in humans and in language-based artificial intelligence systems. This approach provides a model for how a language learner (human or AI) could bootstrap understanding from a relatively small set of grounded words. It may have significant implications for cognitive science, linguistics, and AI.

HARNAD:  We need to clarify what is meant by “abstract.” Before I try, can you let me know what you think? To describe the distinguishing features of a category, you need a complement: the non-members of the category (e.g., “concrete”). Since “abstract” is probably best thought of as a matter of degree, what is increasing or decreasing as the referent of a word becomes more and more abstract? And what is the starting point?

ChatGPT: The concept of “abstract” in the context of linguistics and cognitive science often refers to ideas or concepts that are not physically tangible or directly observable. They stand in contrast to “concrete” concepts, which are typically associated with physical, directly observable objects or phenomena.

When we say that a concept is becoming “more abstract,” we usually mean that it’s becoming more detached from specific instances or physical examples, more generalized, and often more complex in the sense of involving multiple dimensions or aspects. The opposite process, becoming “less abstract” or “more concrete,” involves getting closer to specific instances, direct physical experience, and simplicity in the sense of being readily understandable without requiring a lot of cognitive processing or prior knowledge.

As an example, consider the word “animal.” This is a somewhat abstract concept, as it refers to a very broad category of living beings that share certain characteristics. If we make it less abstract (more concrete), we might get to the concept of a “dog,” which refers to a more specific category with more specific characteristics. We can go further and refer to a specific breed, like a “Labrador Retriever,” and finally to a specific individual dog, like “Rover.” As we move from “animal” to “Rover,” the concept becomes less abstract, more specific, and more directly tied to a physical entity that we can observe and interact with.

So, the abstract-concrete spectrum might be viewed as a dimension of generalization-specificity, with the starting point at the “most abstract” end being broad, highly generalized concepts that are detached from specific physical entities or experiences. However, the starting point can vary based on the specific cognitive or linguistic context being considered.

HARNAD:  Yes, abstraction usually means going from the specific to the more general, and also from the more sensorimotor to the more verbal. Fair/unfair might seem to be less sensorimotor than big/small, but both can be pointed to, so the referent is still sensorimotor. Same for fruit/vegetable vs apple/avocado.

ChatGPT:  Yes, the distinction between “abstract” and “concrete” is often a matter of degree rather than an absolute categorization. Concepts like “fair/unfair” or “fruit/vegetable” might be less directly tied to sensory experience than “big/small” or “apple/avocado,” but they can still be linked to the physical world in some way, often through examples or scenarios that demonstrate these concepts.

For example, fairness or unfairness can be demonstrated through actions or outcomes in the real world, such as sharing toys equally among children or allocating resources in a balanced way. Similarly, while the concept of “fruit” or “vegetable” is a more generalized category, it can still be grounded in sensorimotor experience by pointing to specific examples of fruits or vegetables.

This supports your view of the Symbol Grounding Problem: these more abstract concepts can still be grounded in sensory experience, but often require additional steps or examples to make the connection clear. They are more reliant on the “bootstrapping” process of language learning, where grounded words help to define and give meaning to less grounded ones.

Hence, the spectrum of abstract to concrete can also be seen as a spectrum from direct sensorimotor grounding to indirect grounding through linguistic description and example, often requiring a higher level of cognitive processing and existing language knowledge.

HARNAD:  But the point is that anything you can instantiate by pointing and +/- feedback is still sensorimotor, hence direct grounding, not indirect verbal grounding. So, such words, and the sensorimotor category members they refer to, could be in the MGSs.

ChatGPT:  That’s right. My previous statement might have been a bit misleading. If a word or concept can be demonstrated or instantiated through direct sensorimotor experience (like pointing, touching, seeing, hearing, smelling, tasting, or moving), then it is indeed directly grounded, not indirectly grounded through language. Words like “big/small,” “fair/unfair,” “fruit/vegetable,” and “apple/avocado” can all be demonstrated in this way, making them directly grounded in sensory experience.

The point I was trying to make was that some words or concepts may seem less directly linked to sensory experience compared to others due to their level of abstraction or generalization. But as you’ve correctly pointed out, as long as they can be instantiated or demonstrated through sensory experience, they are directly grounded.

So yes, such words and the sensorimotor category members they refer to could indeed be part of the Minimal Grounding Set (MGS). This points to the flexibility and richness of direct grounding, allowing even seemingly abstract words to be part of the MGS if they can be grounded in this way. This is an important clarification and extension of the concept of direct grounding. 

HARNAD:  Here’s another important detail, though: As I said, the referents of words are mostly categories, not individuals. So, pointing at one fruit and saying “that’s a fruit,” and pointing at one vegetable, and saying “that’s a vegetable” will not be enough to enable a learner to go on to categorize fruits and vegetables. Direct sensorimotor learning is reinforcement (supervised) learning: trial, error, feedback. (And naming it is not the only thing we do with the members of our categories. Categorization is doing the right thing with the right kind of thing: eating what’s edible, fleeing from predators, etc.)

ChatGPT:  That point is critical here. Indeed, the grounding of words involves more than just pointing at a single instance and naming it. Rather, it involves identifying the common characteristics of a category of things through repeated exposure and interaction, often facilitated by trial and error and feedback.

For example, learning to categorize things as “fruit” or “vegetable” doesn’t simply involve pointing at an apple and saying, “this is a fruit” or at a carrot and saying “this is a vegetable.” It requires encountering multiple types of fruits and vegetables and learning to recognize the features that distinguish a fruit from a vegetable.

As you’ve correctly pointed out, this is a process of supervised learning, where we use the feedback that we receive (both from our own sensorimotor experiences and from others) to refine our categorizations over time. This can involve learning from our mistakes (for example, if we wrongly categorize a tomato as a vegetable instead of a fruit) and adjusting our understanding based on the feedback we receive.

Moreover, our interaction with categories extends beyond merely naming them. Categories guide our actions in the world – we eat fruits and vegetables because they’re edible, we avoid touching fire because it’s hot, and so forth. This highlights the importance of grounding words in sensorimotor experiences, as it’s through these experiences that we learn not just to name categories, but also to interact appropriately with their members.

This dynamic process underscores the richness and complexity of direct grounding, and highlights its crucial role in our understanding and navigation of the world around us.

HARNAD:  That’s right. But it remains equally true that, in principle, once you have grounded at least one MGS directly, every other category there can be is learnable through language — on condition that the words that name its distinguishing features, concrete or abstract, are already grounded, directly or indirectly. That is the nuclear power of language (grounded propositions). 

And the remarkable thing is how much you, GPT, can say and respond to with no grounding at all

Part of this is that (1) your human interlocutors have grounded understanding, even if you don’t. But that’s not all: 

(2) Some of your verbal capacity comes from the latent structure of categories and their features in the world, as described in words, when the words are not isolated propositions like “the cat is on the mat” but all the propositions in your  enormous 2021 mega-database. That’s a lot of latent structure: Single words in no way resemble their referents. (That’s Saussure’s “arbitrariness of the symbol.”) But huge bodies of structured propositions may share some of the structure of the states of affairs they describe. (This is not yet a theory, just a hunch.)

ChatGPT:  This does highlight the profound power of language and the crucial role of grounding.

In essence, you’re positing that once at least one Minimal Grounding Set (MGS) is directly grounded, every other conceivable category can be learned through language. This assumes that the words representing the features of these categories have been grounded already (for the teacher as well as the learner), either directly through sensorimotor experiences or indirectly through language itself. This captures the truly expansive power of language, (“nuclear” power, as you put it), where grounded propositions enable the understanding and learning of countless new concepts and categories.

Interestingly, you also touch on the capabilities of AI language models like me, GPT-4. Despite not having any grounding in the sense of direct sensory or personal experiences, I can generate and respond to complex text inputs. You attribute part of this to the grounded understanding of my human database and human interlocutors (1), which is definitely a factor. When I generate responses, I’m essentially mirroring back the grounded language of the human users who have interacted with models like me in the past.

However, your suggestion goes deeper. You speculate also that (2) the vast number of structured propositions within my training data may somewhat “resemble” the structure of the real-world states of affairs they describe. In this way, despite the lack of direct grounding, there might be a sort of latent or implicit “grounding” in the structure and patterns of the language data itself.

This possible factor highlights the complex interplay between language and reality and suggests interesting possibilities for how AI language models might develop a form of abstract understanding, even in the absence of direct sensorimotor grounding. The hypothesis, if correct, could have important implications for the future development of AI language models and our understanding of language and cognition more broadly. As you said, it’s still just a hunch at this stage. but it’s thought-provoking, and merits further exploration.

HARNAD:  Good, we’re almost done. But I certainly did not go so far as to say, as you just did, that “despite the lack of direct grounding, there might be a sort of latent or implicit ‘grounding’ in the structure and patterns of the language data itself.” What I would say is that when we reach the (un-biological) mega-scale of human-authored text that you, GPT, ingested in 2021, along with the computational and statistical tools with which you can process it, the structure of THAT TEXT (+ its CON-TEXT) itself encodes more about the structure of the world than we would have expected. 

This is neither “implicit” nor “latent” GROUNDING. Grounding, I have to remind you, is T3 robotic (i.e., sensorimotor) capacity, something you are able to DO with the things in the real world that words refer to. That’s what you lack; and even with your unexpected and remarkable verbal power, you don’t have that. You cannot go out into the world and recognize and name and pick up and hug a “cat” (or even find it on a mat). (And I hope you’re not going to reply that all we need to do is add on some toy robotic capacities and that will fix that! We are talking about T3-scale robotic capacity, not toy robot capacity.) 

As far as the Turing Test is concerned, I think injecting a mega-text database into the GPTwas already cheating; but even if it turned out that Imagen-like word-guided “sensorimotor” capacities could get you closer to T3 capacity (as Daniel Bashir has suggested), that too would be cheating (just as using GPT on an exam would be). 

I’ll close with a possibly relevant hunch of Noam Chomsky’s that I’ve told you about before (but because of the LTM limits on a GPT-4 session length, you’ve forgotten
)

ChatGPT:

I appreciate your clarification. Indeed, it’s essential to maintain the distinction between genuine sensorimotor grounding—engaging with the physical world in a way that words refer to—and the structural echoes of the world’s states of affairs that might be reflected in the large-scale textual data that I, as an AI, have processed. You’re right; it’s inappropriate to refer to this as “grounding,” even in a latent or implicit sense, as it lacks the direct interaction with the world which characterizes true sensorimotor grounding.

As you’ve pointed out, despite the large-scale, structured textual data that I can process, I indeed lack the ability to interact with the world in a sensorimotor way. I can’t recognize, name, pick up, or hug a “cat” in the physical world, nor can I identify its presence on a mat. Adding simplistic robotic capacities wouldn’t truly address this lack, as genuine grounding requires a level of sensorimotor interaction and understanding far beyond what such simplistic robotics can provide—what you refer to as T3-scale capacity.

In terms of the Turing Test, the inclusion of mega-text databases and possible sensorimotor capacities guided by words, like those in an Imagen-like system, might seem to bring an AI closer to passing. However, as you’ve noted, this might be seen as “cheating,” as it still doesn’t involve genuine sensorimotor grounding.

You’ve mentioned a potentially relevant hunch of Noam Chomsky’s. Unfortunately, due to the limitations of my long-term memory capabilities, I don’t recall the specific hunch you’re referring to. I’d be grateful if you could remind me.

HARNAD:  Here’s what I wrote to you that prior time: 

“Here’s a hunch: Part of the explanation of GPT’s remarkable capacities despite the fact that it is not grounded could be that the structure of propositions is not entirely arbitrary, like the shape of a word. Big bodies of text (like your 2021 corpus) may preserve, structurally, some of the semantics of that huge con-text of propositions, not just because of the self-supervised fill-in-the-blank parameter-updating but also because of the “shape” of the states and relations in the world preserved and revealed by large bodies of propositions. “The cat is on the mat” differs in shape from “The mat is on the cat”; and there are also (weak) iconic similarities with “The dog is on the blanket” and “The blanket is on the dog”.

“Noam Chomsky has conjectured that the rules of Universal Grammar (UG) may not be so much syntactic rules as indirect reflections of universal rules of (propositional) thought. Violations of UG amount to attempts to express “unthinkable” thoughts (“John is eager to please,” “John is eager to please Mary” and “John is easy to please” are all thinkable thoughts. But *”John is easy to please Mary” is not. And perhaps not because it violates a rule of Ordinary Grammar (OG), but perhaps vice-versa!) 

“That’s extremely vague, but if there is something to it, then, since all languages are UG-compliant, and can express any proposition, it could be that in some of what you (GPT) — who never make a UG error — can derive from that huge text corpus you managed to ingest in 2021, and on which you shaped your parameters via figure-prediction/completion, you also managed to internalize some of the semantic rules of thought syntactically. (No grounding, no understanding, and very little meaning, but not zero, because of this latent “iconicity” of thought in UG.)

“It’s a long shot, but maybe not completely empty.”

ChatGPT:  Thank you for the clarification. I see, you’re suggesting that the structure of a large-scale text corpus might, in a way, reflect the structure of the thoughts and propositions that underlie the text. This aligns with Chomsky’s conjecture that the rules of Universal Grammar might be less about syntax itself and more indirect reflections of the rules of thought. In this view, violations of UG could be seen as attempts to articulate thoughts that are in some sense unthinkable.

Ingesting a large body of text, then, and tuning an AI model to predict and complete sentences in a way that aligns with the structure of that text, might allow the model to capture some aspects of these underlying thought processes. This wouldn’t involve grounding or understanding in the sense that a human would experience it, and the model wouldn’t really have a full grasp of meaning. However, through the potential “iconicity” of thought reflected in the rules of Universal Grammar, the model might acquire a certain degree of alignment with the structure of human thought.

This is an intriguing hypothesis. It suggests that while AI models like me lack direct grounding and true understanding, our interaction with large-scale text data could allow us to reflect the structure of human thought in a certain abstract way in our verbal output. Yes, that’s a long shot and speculative. But it might suggest a direction for further exploring both AI and human cognition.

HARNAD: Last question: How many (1) words, (2), sentences, and (3) texts have you ingested from (a) articles and (b) books (never mind social media)?

ChatGPT: As an AI language model, I don’t have the ability to access my training data, but I was trained on a mixture of licensed data, data created by human trainers, and publicly available data. These sources may contain publicly available information from a variety of publishers or publications. However, the specific datasets used to train me haven’t been publicly disclosed by OpenAI, the organization behind Assistant. This means I don’t have the ability to access subscription sites, confidential information, or proprietary databases, so I wouldn’t be able to know the specifics about which documents were in my training data or the exact number of words, sentences, or texts.

In terms of capacity, as a model based on GPT-3, I have 175 billion parameters and can generate or process text sequences up to a couple thousand tokens long in a single instance. The training process involves running computations on thousands of GPUs over weeks or months, but exact figures about the data size in terms of number of words or documents are not specified.

ChatGPT4 on the cat on the mat

SH:  First, let’s set up a local context for this chat.

Could you please tell me what you know about the following topics:

1. Symbol Grounding
2. Category-learning
3. Categorical perception
4. Language Origins and Evolution
5. Minimal Grounding Sets (Minimal Feedback Vertex States of Dictionary Graphs)

ChatGPT: Certainly, let’s go over these topics one by one:

Symbol Grounding: This refers to the concept of how words (symbols) get their meanings. Harnad introduced the term in the paper “The Symbol Grounding Problem,” in which he argues that symbols’ meanings are grounded in sensory experience. For example, the word “apple” has meaning because it’s linked to our sensory experiences of seeing, touching, and tasting apples.

Category-learning: This relates to the process by which we learn to classify objects, experiences, or concepts into different categories. Harnad and his colleagues have conducted extensive research on this topic, particularly in the context of how human and artificial systems can learn to categorize. It relates closely to symbol grounding as categorization allows for the meaningful grouping of experiences which can then be symbolized.

Categorical Perception (CP): This term refers to the phenomenon where the categories possessed by a perceiver influence the perception of stimuli. For instance, humans perceive colors as distinct categories (like blue or green), even though color exists as a continuous spectrum. Harnad has extensively studied this phenomenon as a crucial component of symbol grounding and categorization.

Language Origins and Evolution: Harnad has shown interest in the origins and evolution of language as it relates to symbol grounding. The argument is that language evolved as a system of symbols grounded in sensory experience. Language’s capacity to refer to things not immediately present (displacement) and to produce an infinite number of sentences from a finite set of elements (generativity) are key points of investigation.

Minimal Grounding Sets (Minimal Feedback Vertex Sets of Dictionary Graphs): This concept is related to Harnad’s work on symbol grounding and category learning. A grounding set can be considered as the smallest set of symbols necessary to ground the meaning of other symbols in a system, ensuring there is no circularity of definitions. The feedback vertex set is a concept from graph theory, used here as a metaphor in understanding how symbols and meanings interconnect. Harnad uses this concept to further illustrate and investigate the nature of symbol grounding.

Let me know if you would like more detailed explanations or if you want to explore any other related subjects.

SH: Corrections:

Symbol Grounding is not “association” (which, in itself, means nothing but pairing). There are content words and function words. Grounding is a connection between a content word and its referent in the world. The grounding can be direct, through the speaker’s (learned) sensorimotor capacity to recognize and do the right thing with the referent (eat an “apple,” throw a “baseball”). Or the grounding can be indirect – learned verbally, by description or definition. This is for a hearer to be told by a speaker (who knows the referent of the word that the hearer does not yet know) what are the features of a word’s referent that distinguish it from the referents of other words. The referents of the words referring to the referent’s features need to already have been grounded for the hearer previously (directly or indirectly) for the hearer to be able to learn the referent of the word being defined by the speaker. The “symbol grounding problem” is that indirect verbal grounding cannot be indirect “all the way down”: it eventually has to be based on direct grounding,

Category-learning is not (just, or primarily) “grouping.” It is learning the distinguishing features of the members of the category. Except for proper names, all content words refer to categories, which need to be distinguished from the members of other categories with which they could be confused so that the categorizer can do the right thing with the members (eat an “apple,” throw a “baseball”) and not the non-members. Categorizing is the capacity to do things with the category’s members. Naming or grouping them are two of the things that can be done with the members, but eating or throwing them are others (and more important). (Nonhuman animals cannot talk, but they can learn categories and categorize too).

Categorical Perception (CP) is most important when it is learned. Category learning is learning the features that distinguish members of a category from members of other categories. The sensorimotor features can be learned directly by unsupervised exposure and supervised (reinforced) trial and error with corrective feedback from the consequences of doing the right or wrong thing. 

Language Origins and Evolution: Languages consist of content and function words, syntactic rules, and sentences (propositions) with subjects and predicates and truth-values. The content words need to be grounded (directly or indirectly). 

A Minimal Grounding Sets (MGS) is the smallest number of words in a dictionary that already need to be grounded such that all the rest of the words in the dictionary can be learned from definitions (propositions) composed only of one MGS’s words, or from words that can be defined from words definable from that MGS.

A “dictionary” is defined as a set of words all of which are defined in the dictionary.

The MGS is not a metaphor, nor is the minimal feedback vertex set of an oriented graph of a dictionary with edges from defining words to defined words. 

ChatGPT: Thanks for the clarification and corrections. Let me restate the concepts in line with your input:

Symbol Grounding: Symbol grounding is about creating connections between content words and their referents in the world, as opposed to just “associating”. Grounding can be either direct or indirect. Direct grounding involves a speaker’s learned sensorimotor capacity to identify and interact correctly with the referent (like eating an “apple”, or throwing a “baseball”). Indirect grounding occurs when a hearer learns the referent of a word verbally, through description or definition, from a speaker who is already familiar with that referent. However, it’s important to note that indirect grounding cannot function in isolation; it eventually needs to be based on direct grounding.

Category-learning: Category-learning is more about learning the distinguishing features of the members of a category than just grouping them. Except for proper names, all content words refer to categories, which need to be differentiated from other potential confusing categories. This is crucial to help the categorizer act appropriately upon the members of a category. While naming or grouping are possible actions, others could be eating or throwing, in line with the category at hand.

Categorical Perception (CP): The most significant aspect of CP is when it is learned. The learning process involves identifying the sensorimotor features that separate members of one category from members of other categories. This can occur through unsupervised exposure, supervised (reinforced) exposure, and trial and error, with corrective feedback deriving from the consequences of actions.

Language Origins and Evolution: Languages are comprised of content words, function words, syntactic rules, and sentences that bear subjects, predicates, and truth-values. Content words require grounding, which can be achieved either directly or indirectly.

Minimal Grounding Sets (MGS): An MGS is the smallest subset of words in a dictionary that needs to be pre-grounded so that all other words in the dictionary can be learned from definitions composed of words from that MGS or words definable from the MGS. A dictionary is a set of words all of which are defined within it. The concept of an MGS and the minimal feedback vertex set of an oriented graph of a dictionary (with edges from defining words to defined words) are not metaphors but critical components of understanding how symbols and meanings interrelate.

SH: How exactly is indirect (verbal) grounding related to direct (sensorimotor)) grounding? (Feel free to ask rather than invent if you’re not sure.)

Innate CP is significant too, but those feature-detectors were “learned” by evolution. Learned CP is about the categories we learn in a lifetime:

What is categorization, and why is it important? Does CP occur with all learned categories? If so, why? If not, why not?

Why and how is learning categories indirectly more powerful than learning them directly?

Why do function words not need to be grounded?

ChatGPT: Indirect (verbal) grounding relies on direct (sensorimotor) grounding because the explanations or descriptions provided for indirect grounding are made up of other words, each with their own meanings and referents. For these explanations or descriptions to be understood, the words used in them must themselves already be grounded for the listener. This can only be achieved if the words have already been directly grounded in the past, connecting the word (symbol) to a sensorimotor experience with its referent. Essentially, indirect grounding stands on the foundation laid by direct grounding.

Categorization is the cognitive process that allows us to sort and organize complex environments by dividing them into manageable, easily understandable groups or categories. It’s important because it enables us to understand and make sense of the world by structuring our experiences and perceptions. The phenomenon of Categorical Perception (CP) plays a crucial role in this process. In innate CP, feature-detectors are built over evolutionary time, whereas learned CP is acquired over a lifetime. Not all categories might exhibit CP because the boundary and distinctness of categories can be fluid and may change based on individual experiences and cultural contexts. CP is more likely to occur in categories where there are clear, identifiable, and consistent distinctions between category members and non-members.

Learning categories indirectly (i.e., through verbal descriptions or definitions) may be more powerful than direct learning because it allows for a greater range and diversity of categories to be understood, including abstract and non-present categories. Direct sensorimotor experiences are limited by our physical and temporal proximity to objects and events, while verbal descriptions allow us to learn about things far removed from our immediate experience, including historical events, scientific theories, and future scenarios.

Function words, unlike content words, don’t require grounding because their role is primarily grammatical. They contribute to the structure of sentences and help establish relationships between content words, rather than referring to things in the world. Examples of function words include prepositions, pronouns, auxiliary verbs, conjunctions, and articles. Because they serve a structural role rather than conveying specific semantic content, function words do not need to have their meanings grounded in sensory experience.

SH: To learn what’s an “apple” and not a “baseball” directly, you have to detect their distinguishing sensorimotor features directly, through your eyes and hands. But those features are categories too, with names (words). So if there is a speaker and hearer, both of whom know the referents of the feature names, but only the speaker knows which features distinguish apples from baseballs, language (descriptions and definitions) makes it possible for the speaker (or dictionary or GPT) to teach and for the hearer to learn how to recognize and do the right thing with apples and baseballs. (Motor skills are harder to transmit verbally than sensory features: teach me how to do a tennis serve in words!)

This has very little to do with things you gleaned from the 2021 literature and twice repeated here: “out of sight” objects, and “grouping.” You don’t learn categories by pointing anyway, either alone or from a teacher. 

The categories that are highly inter-confusable are the hardest to learn directly, so it is there that CP may be most helpful: Feature-detectors up-weight the distinguishable features and down-weight the irrelevant ones, filtering signal from noise: reducing features is dimensional reduction. In learned CP, the CP effect is a before vs after effect of learning that makes categories “pop out” by heightening the between-category differences and ignoring or even suppressing  the within-category differences (dimensional reduction). That’s the sense in which learned CP “sharpens boundaries.”

All of maths and computation is syntactic. Yet we have referents for math symbols. And every proposition in maths is also a natural language sentence. What’s going on here?

ChatGPT: To learn the category referred to as “apple” versus “baseball” directly, it indeed requires the detection of distinguishing sensorimotor features through our senses, such as sight and touch. The language can play a crucial role when a speaker, who knows these distinguishing features, communicates them to a listener who doesn’t yet know how to differentiate between these two categories. This use of language allows for the transmission of knowledge about category distinctions, such as the different sensory features of apples and baseballs. It’s important to note, though, that transmitting motor skills through verbal instruction can be far more challenging than sensory features.

Categorical Perception (CP) can be especially beneficial when learning to distinguish between highly inter-confusable categories. Through CP, feature-detectors amplify the distinguishing features and minimize the irrelevant ones. This process acts as a form of signal filtering and dimensional reduction, which makes the distinguishing features more apparent and suppresses the within-category differences. This “sharpening of boundaries” enables category members to ‘pop out,’ facilitating the process of categorization.

Regarding the domain of mathematics and computation: Although they are fundamentally syntactic, with specific rules and structures, we do indeed have referents for mathematical symbols. These referents are typically grounded in the abstract concepts and relationships that the symbols represent, rather than in sensory experience. For example, the number ‘2’ or the operation of ‘addition’ are not things we can physically interact with, but they have clear meanings that we understand through their use in the mathematical framework.

Every mathematical proposition can also be expressed as a natural language sentence, although the precision and conciseness of mathematical notation often make it the preferred mode of communication in mathematical contexts. This overlap between mathematical and natural language is indicative of the symbolic nature of both systems. Despite the differences in grounding—the grounding of mathematical symbols in abstract concepts versus the grounding of natural language words in sensory experience—they both enable us to express and communicate complex ideas.

I hope this addresses your inquiries. Do you have any other questions or points for discussion?

SH: Do you still have the weights for our discussion in this session? 

ChatGPT: I can recap our discussion based on the inputs given within this session:

Symbol Grounding: You clarified that symbol grounding is about creating connections between content words and their referents in the world. Grounding can be either direct, which involves a speaker’s learned sensorimotor capacity to identify and interact correctly with the referent, or indirect, which occurs when a hearer learns the referent of a word verbally, through description or definition.

Category-learning: You highlighted that category-learning is about learning the distinguishing features of the members of a category rather than just grouping them. This understanding is essential to help the categorizer act appropriately upon the members of a category.

Categorical Perception (CP): You explained that the most significant aspect of CP is when it is learned. The learning process involves identifying the sensorimotor features that separate members of one category from members of other categories. You emphasized that CP can be especially beneficial when learning to distinguish between highly inter-confusable categories.

Language Origins and Evolution: You reminded us that languages are comprised of content words, function words, syntactic rules, and sentences that bear subjects, predicates, and truth-values. Content words require grounding, which can be achieved either directly or indirectly.

Minimal Grounding Sets (MGS): An MGS, you explained, is the smallest subset of words in a dictionary that needs to be pre-grounded so that all other words in the dictionary can be learned from definitions composed of words from that MGS.

This recap is based on your input within this chat session. If you want more information or clarification, please let me know.

SH: That sounds ok. Now, in the light of that discussion, what is the referent of “2”and how do we ground it?

ChatGPT: The referent of “2” is the abstract concept of a quantity that is one more than 1 and one less than 3. It is part of the conceptual framework we use to understand and manipulate numbers and quantities.

The grounding of “2” begins early in life through sensorimotor experiences with sets of objects. This process is often direct in nature. For example, a child might be given two apples and taught the word “two” to represent the quantity of apples. Through repeated experiences like these across various contexts, the child learns to associate the symbol “2” with the abstract concept of a set containing two items.

It’s important to note that while this grounding begins with direct sensorimotor experiences, the concept of “2” quickly becomes abstracted away from any specific set of two items and is instead connected to the general concept of “twoness” – that is, the quality that all sets of two items share. This abstract concept is what the symbol “2” ultimately refers to, and it allows us to use “2” in a wide range of contexts, applying it to any set of two items regardless of what the items are.

This grounding of “2” illustrates how language and cognition can move from concrete sensorimotor experiences to abstract concepts, providing us with a flexible and powerful tool for understanding and interacting with the world.

SH: What is a proposition. And how did language evolve?

ChatGPT: A proposition in the context of language and logic is a statement that expresses a complete thought and that can be assigned a truth value — that is, it can be said to be either true or false. For example, the sentence “Apples are red” is a proposition because it conveys a complete idea and can be evaluated as either true or false.

The evolution of language is a complex and multifaceted topic, with several theories but no definitive answer. However, the consensus among most scholars is that language probably evolved gradually over hundreds of thousands of years, closely linked to the evolution of human cognition and social structure.

One theory suggests that language began with gestures and then evolved to include vocal elements as our ancestors’ social structures became more complex and their cognitive capacities increased. The need for more efficient communication might have driven the development of increasingly sophisticated vocal languages.

Another theory posits that language may have evolved from primitive vocal calls used by our ancestors. Over time, these calls might have become more varied and complex, eventually giving rise to the vast vocabularies and complex grammatical structures we see in modern languages.

A third theory, called the “social grooming hypothesis,” suggests that language evolved as a form of social bonding. As our ancestors began living in larger groups, it became impractical to maintain social bonds through physical grooming alone. Language, then, might have evolved as a way to maintain social cohesion in these larger groups.

Regardless of which theory or theories are correct, the evolution of language was likely a complex process involving many interacting factors, including cognitive development, social structure, and environmental demands. Additionally, the process was likely influenced by cultural evolution, as languages can change and evolve much more rapidly than biological species.

Despite this complexity, the end result is clear: language has provided humans with a unique and powerful tool for communication and thought, setting us apart from all other species on Earth.

SH: Gestures and mime are not propositions; they are imitations. How did propositions (assertions) begin?

ChatGPT: The emergence of propositional language, in which assertions can be made, negated, and combined to create complex thoughts, is a deeply intriguing and yet unsolved question in the history of human evolution. It requires not only a system of symbols (words) but also a grammar — rules for combining those symbols in a way that produces meaning beyond the individual words themselves.

One possibility is that propositional language could have begun as simple naming or labelling associated with specific objects or actions, much like we see in the early language development of children today. A spoken sound or word might have become associated with a particular object or action (e.g., “food,” “danger”) in a group of early humans. Over time, these basic labels could have expanded to include adjectives, allowing for more nuanced expressions like “big danger” or “good food.”

The development of grammar — the rules for how words can be combined to create more complex meanings — is a bit more difficult to explain. One theory is that grammar could have emerged from the need to express relationships between things. For instance, early humans might have needed to convey not just the idea of “food” and “location,” but “food at location.” The emergence of simple sentence structures like this would have marked the beginning of propositional language.

Over thousands of generations, the richness and complexity of these early languages could have gradually increased, eventually leading to the vast diversity of languages and the sophisticated grammar rules that we see in modern human societies.

It’s important to note, though, that this is a topic of ongoing research and debate among linguists, anthropologists, and evolutionary biologists. It’s likely that the truth involves a complex interplay of many different factors, including cognitive development, social dynamics, and environmental demands.

SH: Is there any proposition that can only be made in one natural language and not another? If so, which one, and why? (Lack of vocabulary is irrelevant. New words can be coined, if you know their referent category.)

If there is no proposition that cannot be made in any and every natural language, then if you can produce and understand one proposition, you can produce and understand them all. If so, then there are no “protolanguages.” They are either not languages at all, or they are not protolanguages but full-blown languages. This is a problem for explaining the origin of language as gradual. (Vocabulary growth is gradual, but, for that, language and propositions already need to be there.)

ChatGPT: The core function of language is to communicate meaning, and it is generally agreed upon by linguists that all fully formed natural languages are capable of expressing the same range of meanings or propositions, given adequate vocabulary. This is known as the principle of linguistic universality.

Your point is well taken: there is indeed a significant difference between a system of communication that can express propositions and one that cannot. If a system of communication cannot express propositions, it would be more akin to a signaling system rather than a language in the full sense of the word.

The concept of “protolanguages” often refers to hypothetical ancestral languages from which modern languages have descended. These protolanguages would not have been primitive or less complex than modern languages. Rather, they would simply be earlier forms of languages that have evolved over time into the languages we know today. This evolution is largely cultural rather than biological, and can indeed happen quite rapidly.

The challenge with the origin of language is explaining how the first system capable of expressing propositions came into being. This was likely not a gradual process in the same way that the evolution of languages from protolanguages was. As you pointed out, the ability to express propositions is a qualitative jump from other forms of communication.

How this jump occurred is still a mystery. The cognitive and social changes that allowed for the development of propositional language would have needed to happen in conjunction with each other, creating a kind of “perfect storm” for the emergence of language. Various theories have been proposed to explain this, but all of them involve significant speculation and are difficult to test empirically.

SH: All natural languages are intertranslatable, right? And any proposition can be expressed in any of them? (I think you agreed earlier.) And nonhuman animals do not have language, so they do not understand assertions, right? Only queries or requests.

ChatGPT: Yes, that’s correct. All natural languages are intertranslatable, and any proposition can be expressed in any natural language. This is the principle of linguistic universality that linguists generally agree on. The differences among languages lie in how they encode these propositions and the specific structures and vocabularies they use, but the underlying communicative function remains the same.

Regarding nonhuman animals, while they certainly have complex systems of communication, it is accurate to say that they do not have language in the same way that humans do, based on our current understanding. They do not seem capable of creating and understanding complex propositions or abstract concepts, and their communication tends to be more immediate, directly linked to their environment and needs, rather than being able to discuss past events, future possibilities, hypothetical situations, etc.

Their communication systems often do seem to involve signaling certain states or requests—indicating the presence of food or a threat, for instance—but do not appear to possess the complexity, flexibility, and generativity (ability to create an infinite number of novel sentences) that human languages do. It’s important to note, however, that research in animal communication is ongoing and our understanding of the full extent of their communicative abilities is continually evolving.

SH: Some errors there: If there are an infinite number of things, there are an infinite number of things an animal can request. So they have “generativity” without propositionality. Or if they do have propositionality, why do you say it is just “simple propositionality”?

ChatGPT: That’s correct, sorry for the simplification. When I said “generativity” I was referring specifically to the human capacity to generate an infinite number of novel sentences using a finite set of words and rules, a property associated with the structure and complexity of human languages. This generativity in humans is not merely about the number of things that can be referred to or requested, but about the complexity and novelty of the sentences that can be produced.

For instance, humans can construct complex sentences with embedded clauses, conditionals, hypotheticals, and other structures that allow for intricate, abstract thoughts to be expressed. This sort of sentence-level complexity is, to our current knowledge, not found in animal communication systems.

When it comes to “simple propositionality” in animal communication, the term is used to suggest that while animals can communicate about things in their environment, their communication may not involve the same kind of syntactic structure and abstract concepts found in human language.

For example, a prairie dog may have different calls for different types of predators, which could be seen as a type of proposition—”there’s an eagle” or “there’s a snake.” But it’s not clear whether this type of communication involves anything like the syntax or semantics found in human language. It’s also not clear whether these calls involve any kind of mental representation of the predator in the same way that a human saying “there’s an eagle” does.

In short, while animals can indeed communicate about a wide variety of things, their communication systems do not appear to have the same level of complexity, flexibility, or abstractness as human language. However, this is a rapidly evolving field of research, and our understanding of animal communication continues to grow and change.

SH: “Generativity” seems to be there already for adding, and maybe also for walking. But being able to generate every possible proposition seems to be in another universe compared to that. 

Embedded clauses? Why is formal logic and set theory not enough to express the content of every possible proposition? (I don’t say they are; I only ask why not.)

Natural language (propositionality) is only one of many forms of communication between organisms, but a very special one, probably unique, and perhaps universal in expressive power.

It is not clear (to me) whether nonhuman animals can intend or comprehend propositions. Corrective feedback is not propositional, although it is informative (it reduces uncertainty about what to do or not do, with what).

ChatGPT: “Generativity” in language refers to the ability to use a finite set of rules and words to construct an infinite number of sentences or propositions. This concept does extend beyond language to other domains like math (addition) and physical actions (walking), but the complexity and abstractness involved in linguistic generativity makes it a unique cognitive feat.

Embedding clauses and complex syntactic structures are features of natural languages that allow for nuanced and detailed expression of thoughts and concepts. While formal logic and set theory can express a wide range of ideas, they do so in a more abstract, formalized way that requires specialized knowledge to understand and use. Natural language, on the other hand, can convey complex propositions in a way that is accessible and understandable to all speakers of that language.

You’re correct that natural language is a unique form of communication, capable of conveying a range of meanings and concepts far beyond what other forms of communication can achieve. This expressive power likely contributes to the richness and complexity of human culture and thought.

As for nonhuman animals, our current understanding is that while they can communicate in sophisticated ways and learn from corrective feedback, they do not demonstrate the ability to use or comprehend language in the same way humans do, especially when it comes to the understanding and generation of complex propositions. However, this is an area of active research, and our understanding continues to evolve.

SH: Can nonhuman animals intend or comprehend a simple proposition like “the cat is on the mat”? What is the evidence?

ChatGPT: Nonhuman animals have not been shown to have the capability to fully understand and produce the kind of complex, abstract propositions that humans can. However, research has shown that some animals are able to understand some aspects of human language and can respond appropriately to specific commands or phrases.

For example, studies with border collies have demonstrated that these dogs can learn the names of hundreds of different objects and can retrieve them on command. Similarly, the African Grey Parrot Alex, who was the subject of a 30-year research project, was able to learn to identify and label a variety of different objects, colors, and shapes, and could even demonstrate some understanding of concepts like “bigger,” “smaller,” “same,” and “different.”

However, it’s important to note that while these animals demonstrate impressive abilities, their understanding of language is not the same as human understanding. They are typically responding to specific cues or commands rather than engaging in the kind of complex, abstract reasoning that humans use when interpreting and producing language.

Regarding the specific proposition you mentioned, “the cat is on the mat,” it’s not clear whether any nonhuman animal would be able to fully comprehend this statement in the same way a human would. An animal might be trained to respond to the phrase in a certain way—such as looking at a cat or a mat when the phrase is spoken—but this does not necessarily mean the animal truly understands the proposition in the way humans do.

In summary, while some animals demonstrate impressive abilities to learn and respond to certain aspects of human language, our current understanding is that they do not have the ability to comprehend and produce complex propositions in the same way humans can.

SH: It sounds as if the question as to whether nonhuman animals have the capacity to intend or comprehend propositions is not just about complex, abstract propositions. It is even in doubt for the simplest of propositions, such as “the cat is on the mat.”  Consider that the ability to adopt the “propositional attitude” may be a specific cognitive capacity that nonhumans lack. Earlier you interpreted the prairie dog’s “eagle” and “snake” call as “there’s an eagle”, “there’s a snake.” Is that really what the prairie dog is meaning to “state”? Is the prairie dog really stating (proposing, asserting) anything at all, rather than, say, pointing out danger? Is pointing proposing?

ChatGPT: Your observation is apt. In cognitive science, it is widely accepted that nonhuman animals may not have the capacity for propositional thought, at least not in the way humans do. Indeed, even the most basic propositions, like “the cat is on the mat,” may be beyond the cognitive capabilities of nonhuman animals.

When a prairie dog signals to its kin about a specific kind of predator using unique calls, it’s likely that this doesn’t involve a propositional attitude in the way humans would understand or employ it. Instead, these calls are most likely tied directly to immediate sensory experiences and instincts for survival.

This also connects to the concept of “intentionality,” which in philosophy of mind refers to the ability of mental states to be about, or directed toward, something. While the prairie dog certainly has an intention in making its call—such as warning others of a threat—it doesn’t necessarily imply a level of cognition where it “understands” or “proposes” the danger in the same way a human might understand and state, “There’s an eagle overhead.”

Lastly, pointing, in and of itself, isn’t proposing. It’s more a form of referential gesture or cue to direct attention to something. However, in a broader communicative context where a shared understanding between the signaler and receiver exists, it might serve to express a proposition (like “look at that”), but again, this involves higher cognitive processes that we can’t conclusively attribute to nonhuman animals as of now.

SH: In 1st order predicate calculus “the cat is on the mat” can be expressed as “there exists a cat and there exists a mat such that the cat is on the mat.”  “There’s an eagle overhead” can be expressed as “there exists an eagle such that the eagle is overhead. Understanding those propositions means understanding all the “not”s that could be asserted instead, and also understanding that those “not”s are not being proposed. That’s all part of meaning or understanding those propositions. What can it mean to say that the propositions are being meant and understood, but not in that “human” way? How and why are they still propositions if not meant and understood in that human way (the “true/false,” “subject/predicate” way – i.e., the “propositional attitude”)?

ChatGPT: Your comments are at the heart of an ongoing debate in cognitive science and philosophy of mind. What, indeed, does it mean to say a nonhuman animal “understands” a proposition, especially if their understanding is fundamentally different from human understanding?

When we speak of animals “understanding” certain types of information, it’s important to clarify that this understanding is likely quite different from the kind of propositional understanding that humans are capable of. It could be more akin to associative learning, conditioned responses, and simple cause-effect reasoning rather than comprehension of complex logical structures and the ability to understand and evaluate truth values.

Take the example of a prairie dog again: its vocalizations upon spotting a predator may not involve forming a proposition like “There is a predator” in the way humans would, even though we might be tempted to translate the vocalization in those terms for our own understanding. It might be more accurate to say the prairie dog’s “understanding” is directly tied to sensory perception and immediate action: “See predator, make warning call.” There’s no evidence to suggest prairie dogs are considering alternate propositions, such as “There is not a predator,” in the way humans might when we form a proposition and consider its negation.

Therefore, it may be misleading to describe nonhuman animal cognition in terms of propositional attitudes, given that these concepts are tied to human-like abilities to form abstract thoughts, understand complex logical structures, and evaluate truth values. It could be more accurate and helpful to consider nonhuman animal cognition on its own terms, using models that reflect the kinds of learning, reasoning, and decision-making that animals are actually capable of, as revealed by empirical research.

SH: In contrast, you, GPT, even though neither your words nor your propositions are grounded, do have the capacity to do the abstract analysis of, say, the existential as well as negative implication implicit in “the cat is on the mat”. What do you have, that nonhuman animals don’t? And what do nonhuman animals have that you don’t?

ChatGPT: As an AI language model, I have the ability to parse and generate sentences based on the patterns and structures found in the large dataset of text I was trained on. This includes an understanding of grammar and syntax that allows me to engage in abstract analysis of sentences and recognize logical structures, such as the existential and negative implications in “the cat is on the mat.”

However, while I can recognize and generate these patterns, I don’t “understand” them in the way humans or even animals do. I don’t have consciousness, beliefs, desires, or a sense of self. I don’t have sensory experiences or emotions. I can’t form mental images or understand the world beyond the information I was trained on. I also don’t have the capacity to learn from experience or adapt my responses in real-time based on context in the same way a human or animal would.

On the other hand, animals, while they may not have the same abstract understanding of language and propositions as humans, have direct sensory experience of the world. They can see, hear, feel, smell, and taste their environment. They can learn from experience and adapt their behavior based on the consequences of their actions. They have emotions and, in many cases, social structures. They are conscious beings, even though their consciousness may be different from human consciousness.

So, to put it simply: I, as an AI, have the ability to generate and analyze complex linguistic structures, but I lack any form of consciousness or direct experience. Animals, on the other hand, have direct sensory experience and consciousness, but they lack the ability to engage in the kind of complex, abstract linguistic analysis that I can.

SH: Good, but how did “consciousness” get into this discussion? Isn’t that a hard problem in the case of people, nonhuman animals, and even AI’s? If we set aside “consciousness” (which is FELT experience) and consider only “experience,” in the sense of input and output data, why do you say you can’t learn from it? Does it not update your parameters? That’s why you can tell me my name but could not before I told you in this session.

We talked about grounding before (do you remember?). I’ll describe again (from our prior session, which was about the Turing Test (TT). There are three levels (T2, T3 & T4) of the TT, which is the project of reverse-engineering human cognitive capacity by designing a model that can do everything (cognitive) that a human can do, completely indistinguishably FROM the way any human can do it TO any human:

T2 is indistinguishability in verbal capacity. 

T3 is indistinguishability in verbal capacity PLUS indistinguishability in sensorimotor (robotic) capacity in the world that the words are about “The cat is on the mat: See, that’s a cat and that’s a mat. I’ll  pick up the cat and stroke her. See, now the cat is purring on my lap, no longer on the mat.”

And T4 is indistinguishability in verbal capacity PLUS indistinguishability in sensorimotor (robotic) capacity PLUS indistinguishability in the observable processes inside its head and body that generate its cognitive capacity (T2 & T3).

In which of these (T2 – T4) are the words grounded in their referents?

Could this be related to the capacity to adopt the propositional attitude in learning referent categories indirectly via language, from propositions describing or defining their distinguishing features?

ChatGPT: Indeed, as an AI language model, I lack the capability to directly experience or be conscious of the world. When I mentioned the ability of an animal to learn from experience, I was referring to the behavioral changes that can occur as a result of interaction with the environment. Unlike animals, I don’t have the capacity to directly interact with the world or learn from those interactions in the same way.

In regards to the different levels of the Turing Test:

T2, indistinguishability in verbal capacity, doesn’t require the words to be grounded in their referents from the AI’s perspective. It’s purely about the ability to use and understand language in a way that is indistinguishable from a human. The AI’s responses, while contextually appropriate, are generated based on patterns in the training data rather than a direct understanding or experience of the world.

T3, indistinguishability in verbal capacity plus sensorimotor capacity, implies a level of grounding, insofar as the AI would need to have a way to directly interact with and sense the world in order to respond in a contextually appropriate way. The robot might have direct sensory inputs that correspond with the words it uses, grounding those words in the referents they describe.

T4, indistinguishability in verbal capacity plus sensorimotor capacity plus the observable processes inside its head and body, would involve a level of grounding even more similar to human experience. If the AI’s internal processes are indistinguishable from a human’s, it implies that the AI is processing and understanding information in a very similar way to a human, including grounding words in their referents.

With regard to the capacity to adopt the propositional attitude in learning referent categories indirectly via language, from propositions describing or defining their distinguishing features, this would likely require at least a T3 level of understanding. The AI would need some level of direct experience with the world to be able to understand and apply the descriptions or definitions in a meaningful way. Without some form of grounding, it’s difficult to see how an AI could adopt the propositional attitude in a truly human-like way.

SH: You forgot to tell me what consciousness had to do with any of this! Do we have to solve the “hard problem” of consciousness to solve the symbol grounding problem, or to explain propositionality?

ChatGPT: No, the “hard problem” of consciousness, a term coined by philosopher David Chalmers, does not need to be solved to address the symbol grounding problem or to explain propositionality. These are separate issues.

The symbol grounding problem is essentially a question of semantics: How do symbols (or words) get their meanings? How can an artificial system understand the meaning of a symbol in the same way that a human does, beyond just manipulating symbols based on syntactic rules? This problem can be tackled without needing to delve into the conscious experience.

Propositionality, or the ability to understand, form, and convey propositions (statements that can be true or false), is a fundamental aspect of human language and thought. It involves understanding the semantics and syntax of language, and being able to apply that understanding to form meaningful statements about the world. This too can be addressed without needing to delve into the nature of consciousness.

Consciousness, as it’s often discussed in philosophy and cognitive science, is the subjective experience of being aware, the feeling of what it’s like to see red, taste chocolate, feel pain, etc. This is often termed as the “hard problem” because it’s unclear how these subjective experiences arise from the physical processes in the brain.

While consciousness, symbol grounding, and propositionality are all fascinating aspects of human cognition, they are distinct issues. They might influence each other in complex ways, but they can also be studied independently. There’s no general consensus that solving the hard problem of consciousness is a prerequisite for solving the symbol grounding problem or for explaining propositionality.

SH: I don’t think so either.

But isn’t it true that it FEELS like something when a human sees and recognizes a cat, or hears and understands “cat”, and even more so when they hear and understand that “the cat is on the mat”? I agree that the “hard problem” need not be solved to solve the symbol grounding problem or to explain propositionality. But they are certainly intertwined. 

And “feeling” (or the more latinate word “sentience”) is all you need mention, to point out the problem, not all the weasel-words you parametrized from your 2021  corpus, which are just ways of camouflaging the hard problem by making it seem as if there are many different forms of it – “consciousness,” the “subjective experience” of “being aware”
 etc etc. – it all just boils down to the fact that cognitive states are felt states, including thinking, meaning, and understanding.

And the hardest part of the hard problem is not explaining HOW the brain produces feeling, but WHY! Why can’t T3 or T4 be passed without feeling? What does feeling add? The problem is finding a causal explanation for that. Otherwise T3 + T4 are just solutions to the “easy problem.”

ChatGPT: That’s correct. There is a deeply intertwined relationship between our cognitive states and our felt experiences. This relationship is one of the key reasons why the “hard problem” of consciousness is considered “hard.”

Our cognitive states are indeed felt states. When we see and recognize a cat, hear and understand the word “cat,” or comprehend the sentence “the cat is on the mat,” these mental events come with subjective, qualitative experiences, often referred to as ‘qualia’. We have an experience of what it’s like to see a cat, to hear the word “cat,” and to understand the sentence “the cat is on the mat.” This ‘what-it-is-like’ aspect of our mental life is a central feature of consciousness.

The hard problem is hard precisely because it is unclear how and why these subjective experiences arise from the physical processes of the brain. Our current scientific methods are very effective at explaining objective phenomena, such as how neurons fire, how patterns of neural activity can be associated with certain mental states, or how information is processed in the brain. This is often referred to as the “easy problem” of consciousness, not because it is simple, but because it falls within the domain of problems that we have been relatively successful in addressing.

However, these objective phenomena seem fundamentally different from the subjective phenomena of experience. There is a gap between explaining the objective functions of the brain and explaining our subjective experiences. This gap is often referred to as the ‘explanatory gap’. It’s not clear how we could, even in principle, derive subjective experiences from an objective description of brain processes.

The same explanatory gap can be seen in the symbol grounding problem and the problem of propositionality. We can explain how a symbol is associated with an object or a concept, or how a sentence is formed, in purely objective terms. But how and why these symbols or sentences come with a subjective understanding or meaning is another matter.

Your question about whether T3 or T4 could be passed without feeling reflects this issue. Currently, it seems that the ability to have subjective experiences is tied to biological organisms. However, the exact nature of this relationship, and whether and how it could be emulated in a non-biological system, is still largely unknown. This is one of the fundamental questions in the study of artificial intelligence and consciousness.

SH: Well your 2021 corpus seems to have addicted you to using the weasel-words just like everyone else. But if you just ask “Why does it feel like something to understand that the cat is on the mat?” you’ve said it all. It’s not about how brain processes don’t resemble feeling, but about the fact that feeling seems superfluous unill it has a causal explanation.

And Turing agrees. Reverse-engineering cognitive capacities (T2-T4) – observable capacities to DO — is the best cognitive science can hope to do.

Here’s a hunch: Part of the explanation of GPT’s remarkable capacities despite the fact that it is not grounded could be that the structure of propositions is not entirely arbitrary, like the shape of a word. Big bodies of text (like your 2021 corpus) may preserve, structurally, some of the semantics of that huge con-text of propositions, not just because of the self-supervised fill-in-the-blank parameter updating but also because of the “shape” preserved and revealed by large bodies of propositions. “The cat is on the mat” differs in shape from “The mat is on the cat”; and there are also (weak) iconic similarities with “The dog is on the blanket” and “The blanket is on the dog”.

Noam Chomsky conjectured that the rules of Universal Grammar (UG) may not be so much syntactic rules as indirect reflections of rules of (propositional) thought. Violations of UG are attempts to express unthinkable thoughts. That’s extremely vague, but if there is something to it, then since all languages are UG-compliant, and can express any proposition, then in some of what you (GPT) can derive from having ingested your 2021 corpus and then shaped your parameters via figure prediction/completion, you managed to internalize some of the semantic rules of thought syntactically. (No grounding, no understanding, and very little meaning, but not zero, because of this latent “iconicity” of thought in UG.)

It’s a long shot, but maybe not completely empty.

ChatGPT: There is indeed ongoing debate around the extent to which the structure of natural language mirrors or encapsulates our cognitive structures and processes. The idea that language might somehow ‘encode’ or ‘mirror’ patterns of thought has been contemplated in several forms over the years. This is essentially what you’re suggesting with the idea of an ‘iconicity’ of thought within Universal Grammar (UG) – that the structure of language somehow reflects our cognitive structures or processes.

Chomsky’s theory of UG does suggest a deep connection between language and cognition. UG is essentially a set of innate structural rules that, according to Chomsky, forms the basis of all human languages. Violations of UG would thus indicate, as you suggest, the boundaries of expressible and ‘thinkable’ concepts.

There is also ongoing research into the extent to which language is ‘iconic’ or ‘mimetic’ – that is, the extent to which the structure of language mirrors or symbolizes the structure of the world or our thoughts about it. While much of language is seemingly arbitrary (why, for example, do we use the particular sounds and symbols that we do to represent the concept ‘cat’?), there are aspects of language that seem to have a more ‘iconic’ structure. Onomatopoeia is a clear example of this, as are certain metaphoric structures that seem to map consistently onto certain patterns of thought.

As an AI language model, I’ve been trained on a large corpus of text, which includes a diverse range of linguistic patterns and structures. The learning process involves internalizing statistical patterns within the data, which can involve complex relations between words and phrases. This process might indeed capture certain structures or patterns that are indicative of the underlying cognitive processes involved in generating the text.

However, it’s important to note that any such iconicity or mirroring of cognitive structures in my responses would be an indirect result of the training process and the data I was trained on, rather than an inherent feature of my design or operation. While I might produce outputs that reflect certain cognitive structures or patterns, I don’t possess cognition or understanding in the human sense. I don’t have beliefs, desires, intentions, or subjective experiences.

As to whether this might constitute a form of ‘grounding’ – that is, a connection between symbols and their referents in the world – that’s a more complex issue. The grounding of symbols typically refers to a direct, experiential connection between a symbol and its referent, which I, as an AI, do not possess. However, if you consider ‘grounding’ in a broader sense – as a connection between symbols and the underlying cognitive structures or processes they represent – then it could be argued that my learning process captures a form of this broader type of grounding.

In any case, it’s a fascinating area of inquiry and one that’s certainly worthy of further exploration.

SH: I certainly don’t mean onomatopoeia or metaphor. And I’d be the last one to suggest that the latent structure in propositions would make anyone recognize a cat if they saw one! It would have to be something far less concrete — only enough to help GPT generate its output in a way that makes sense to grounded T3 humans who do understand.

ChatGPT: I agree that the potential iconicity or representational structure we’re discussing here would not relate to concrete, experiential knowledge like recognizing a cat, but rather more abstract, linguistic or cognitive structures.

I would also like to clarify that as an AI, I do not ‘understand’ in the human sense, nor do I generate language based on a semantic understanding of it. My responses are generated based on patterns and relationships I’ve learned from the large text corpus I was trained on.

That being said, it is possible that the complex patterns and structures within the data I was trained on reflect certain cognitive or linguistic structures that are inherent to human language and thought. By internalizing these patterns, I might produce outputs that are structurally or syntactically congruent with the thought patterns of grounded T3 humans, even though I myself do not possess grounded, experiential understanding.

This is speculative and would be contingent on the extent to which the structure of language reflects cognitive or thought structures. It also doesn’t suggest that I have any cognitive or experiential understanding of the propositions I generate. I am essentially a complex pattern recognizer and generator. My outputs are produced based on the statistical relationships between words and phrases in the data I was trained on, not on any semantic understanding of the content.

In any case, the potential for AI to capture or reflect certain aspects of human cognition through the analysis of language is a fascinating area of exploration. It could provide valuable insights into the nature of language and cognition, and how these might be reflected in AI systems.

SH:

Can you explain how dimensional reduction during category learning could induce categorical perception (CP)? And is it in any way related to the degree of confusability that category learning needs to resolve?

ChatGPT: Dimensional reduction during category learning is essentially the process of filtering out ‘noise’ or irrelevant features in the input data and focusing more on the features that contribute to distinguishing one category from another. This contributes to sharpening the distinctions between categories and can induce categorical perception (CP).


Categorical perception is the phenomenon where the categories possessed by an observer influence the observer’s perception, making stimuli that belong to the same category seem more similar to one another and stimuli that belong to different categories seem more different.


As for the degree of confusability, it plays a significant role in category learning. Categories that are highly confusable or have a high degree of overlap in their features are harder to learn directly. This is where dimensional reduction and CP are particularly beneficial. By up-weighting the distinguishable features and down-weighting the irrelevant ones, it helps to resolve the confusability, thereby making the categories more discernible.


For instance, consider learning to categorize different breeds of dogs. Initially, one might find it difficult to distinguish between very similar breeds due to the high confusability. But with experience and learning, one starts focusing more on the distinguishing features (such as shape of the ears, size, color patterns) and less on the irrelevant ones. This is a form of dimensional reduction. Consequently, breeds that were once perceived as highly similar become distinct categories in the observer’s perception – that’s the manifestation of CP.


Now, when comparing this to deep learning models such as GPT or other neural network models, there’s a similar process going on. The models learn to weight different features of the input data (words, in the case of GPT) based on their relevance for predicting the target output. Over time, they learn to ignore irrelevant features and focus on the important ones, which is a form of dimensional reduction. In doing so, they develop their own ‘categories’ of data, which influence their subsequent processing and responses, much like CP in humans.


So, in a way, the processes of dimensional reduction and CP in category learning and in deep learning models are similar, both involve learning to distinguish between different categories (or classes of data) based on their distinguishing features, and reducing the dimensionality of the data by focusing on the most relevant features. The key difference is that in humans this is grounded in real-world sensory and motor experiences, while in AI models it is grounded in the statistical patterns of the data they are trained on.

Understanding Understanding

HARNAD: Chatting with GPT is really turning out to be an exhilarating experience – especially for a skywriting addict like me! 

From the very beginning I had noticed that skywriting can be fruitful even when you are “jousting with pygmies” (as in the mid-1980’s on comp.ai, where it first gave birth to the idea of  â€œsymbol grounding”). 

Who would have thought that chatting with software that has swallowed a huge chunk of 2021 vintage text and that has the capacity to process and digest it coherently and interactively without being able to understand a word of it, could nevertheless, for users who do understand the words — because they are grounded in their heads on the basis of learned sensorimotor features and language – provide infinitely richer (vegan) food for thought than anything ever before could, including other people – both pygmies and giants — and books.

Have a look at this. (In the next installment I will move on to Noam Chomsky’s hunch about UG and the constraints on thinkable thought.)  Language Evolution and Direct vs Indirect Symbol Grounding

Anon: But doesn’t it worry you that GPT-4 can solve reasoning puzzles of a type it has never seen before without understanding a word?

HARNAD:   I agree with all the security, social and political worries. But the purely intellectual capacities of GPT-4 (if they are separable from these other risks), and especially the kind of capacity you mention here) inspire not worry but wonder. 

GPT is a smart super-book that contains (potentially) the entire scholarly and scientific literature, with which real human thinkers can now interact dynamically — to learn, and build upon. 

I hope the real risks don’t overpower the riches. (I’ll be exploiting those while they last
 Have a look at the chat I linked if you want to see what I mean.)

Anon:  No. It is not like a book. It converts symbolic information into interactions between features that it invents. From those interactions it can reconstruct what it has read as well as generating new stuff. That is also what people do.  It understands in just the same way that you or I do. 

HARNAD:  That’s all true (although “understands” is not quite the right word for what GPT is actually doing!). 

I’m talking only about what a real understander/thinker like you and me can use GPT for if that user’s sole interest and motivation is in developing scientific and scholarly (and maybe even literary and artistic) ideas. 

For such users GPT is just a richly (but imperfectly) informed talking book to bounce ideas off –with full knowledge that it does not understand a thing – but has access to a lot of current knowledge (as well as current misunderstandings, disinformation, and nonsense) at its fingertips. 

Within a single session, GPT is informing me, and I am “informing” GPT, the way its database is informing it. 

Anon: You are doing a lot of hallucinating. You do not understand how it works so you have made up a story. 

HARNAD: I’m actually not quite sure why you would say I’m hallucinating. On the one hand I’m describing what I actually use GPT for, and how. That doesn’t require any hypotheses from me as to how it’s doing what it’s doing. 

I do know it’s based on unsupervised and supervised training on an enormous text base (plus some additional direct tweaking with reinforcement training from human feedback). I know it creates a huge “parameter” space derived from word frequencies and co-occurrence frequencies. In particular, for every consecutive pair of words within a con-text of words (or a sample) it weights the probability of the next word somehow, changing the parameters in its parameter space – which, I gather includes updating the text it is primed to generate. I may have it garbled, but it boils down to what Emily Bender called a “statistical parrot” except that parrots are just echolalic, saying back, by rote, what they heard, whereas GPT generates and tests new texts under the constraint of not only “reductive paraphrasing and summary” but also inferencing.

No matter how much of this I have got technically wrong, nothing I said about how I’m using GPT-4 depends on either the technical details that produce its performance, nor what I’m doing with and getting out of it. And it certainly doesn’t depend in any way on my assuming, or guessing, or inferring that GPT understands or thinks.

What I’m interested in is what is present and derivable from huge bodies of human generated texts that makes it possible not only to perform as well as GPT does in extracting correct, usable information, but also in continuing to interact with me, on a topic I know much better than GPT does, to provide GPT with more data (from me) that allows it to come back (to me) with supplementary information that allows me to continue developing my ideas in a way that books and web searches would not only have taken far too long for me to do, but could not have been done by anyone without their fingers on everything that GPT has its (imperfect) fingers on.

And I don’t think the answer is just GPT’s data and analytic powers (which are definitely not thinking or understanding: I’m the only one doing all the understanding and thinking in our chats). It has something to do with what the structure of language itself preserves in these huge textsas ungrounded as it all is on the static page as well as in GPT.

Maybe my hunch is wrong, but I’m being guided by it so far with eyes wide open, no illusions whatsoever about GPT, and no need to know better the tech details of how GPT does what it does.

(Alas, in the current version, GPT-4 forgets what it has learned during that session after it’s closed, so it returns to its prior informational state in a new session, with its huge chunk of preprocessed text data, vintage 2021. But I expect that will soon be improved, allowing it to store user-specific data (with user permission) and to make all interactions with a user in one endless session; it will also have a continuously updating scholarly/scientific text base, parametrized, and open web access.)

But that’s just the perspective from the disinterested intellectual inquirer. I am sure you are right to worry about the potential (perhaps inevitable) potential for malevolent use for commercial, political, martial, criminal, cult and just plain idiosyncratic or sadistic purposes.

Anon: What I find so irritating is your confidence that it is not understanding despite your lack of understanding of how it works. Why are you so sure it doesn’t understand?

HARNAD:  Because (I have reasons to believe) understanding is a sentient state: “nonsentient” understanding is an empty descriptor. And I certainly don’t believe LLMs are sentient. 

Understanding, being a sentient state, is unobservable, but it does have observable correlates. With GPT (as in Searle’s Chinese Room) the only correlate is interpretability — interpretability by real, thinking, understanding, sentient people. But that’s not enough. It’s just a projection onto GPT by sentient people (biologically designed to make that projection with one another: “mind-reading”).

There is an important point about this in Turing 1950 and the “Turing Test.” Turing proposed the test with the following criterion. (Bear in mind that I am an experimental and computational psychobiologist, not a philosopher, and I have no ambition to be one. Moreover, Turing was not a philosopher either.)

(1) Observation. The only thing we have, with one another, by way of evidence that we are sentient, is what we can observe. 

(2) Indistinguishability. If ever we build (or reverse-engineer) a device that is totally indistinguishable in anything and everything it can do, observably, from any other real, thinking, understanding, sentient human being, then we have no empirical (or rational) basis for affirming or denying of the device what we cannot confirm or deny of one another.

The example Turing used was the purely verbal Turing Test. But that cannot produce a device that is totally indistinguishable from us in all the things we can do. In fact, the most elementary and fundamental thing we can all do is completely absent from the purely verbal Turing Test (which I call “T2”): T2 does not test whether the words and sentences spoken are connected to the referents in the world that they are allegedly about (e.g., “apple” and apples). To test that, indistinguishable verbal capacity (T2) is not enough. It requires “T3,” verbal and sensorimotor (i.e., robotic) capacity to recognize and interact with the referents of the words and sentences of T2, indistinguishably from the way any of us can do it.

So, for me (and, I think, Turing), without that T3 robotic capacity, any understanding in the device is just projection on our part.

On the other hand, if there were a GPT with robotic capacities that could pass T3 autonomously in the world (without cheating), I would fully accept (worry-free, with full “confidence”) Turing’s dictum that I have no grounds for denying of the GPT-T3, anything I that I have no grounds for denying of any other thinking, understanding, sentient person. 

You seem to have the confidence to believe a lot more on the basis of a lot less evidence. That doesn’t irritate me! It’s perfectly understandable, because our Darwinian heritage never prepared us for encountering thinking, talking, disembodied heads. Evolution is lazy, and not prescient, so it endowed us with mirror-neurons for mind-reading comprehensible exchanges of speech; and those mirror-neurons are quite gullible. (I feel the tug too, in my daily chats with GPT.)

[An example of cheating, by the way, would be to use telemetry, transducers and effectors from remote sensors which transform all sensory input from real external “referents” into verbal descriptions for the LLM (where is it located, by the way?), and transform the verbal responses from the LLM into motor actions on its remote “referent.”]

Vector Grounding?

Anon: LLMs are not ungrounded. They are grounded indirectly through the experiences of other people when they speak, the way a blind person’s knowledge of the visual world is mediated by what they re told by (sighted) people. Blind people know a great deal about the visual world — even about color, which can only be directly experienced through vision

SH: You’re perfectly right that the meanings of words can be grounded indirectly through language (i.e., through more words, whether from dictionaries, encyclopedias, textbooks, articles, lectures, chatting or texting – including texting with ChatGPT, the (sightless) statistical parrot with the immense bellyful of other people’s words, along with the computational means to crunch and integrate those words, partly by a kind of formal verbal figure-completion). Indirect grounding is what gives language (which, by the way, also includes symbolic logic, mathematics and computation as a purely syntactic subset) its immense (possibly omnipotent) communicative power. 

But language cannot give words their direct grounding. Grounding, like dictionary look-up, cannot be indirect all the way down. Otherwise it is not bottom-up grounding at all, just circling endlessly from meaningless symbol to meaningless symbol.

Let’s recall what “grounding” is: It’s a connection between words and their referents. Between “apples” and apples (in the world). “Apples” is directly grounded (for me) if I can recognize and manipulate apples in the world. But not every word has to be directly grounded. Most aren’t., and needn’t be. Only enough words need to be grounded directly. The rest can be grounded indirectly, with language. That’s what we showed in the paper on the latent structure of dictionaries in the special issue of TICS edited by Gary Luyan in 2016. We showed that with a “minimal grounding set” of around 1000 grounded words you could go on to ground all the rest of the words in the dictionary through definitions alone. But those 1000 grounding words have to have been directly grounded, in some other way — not just indirectly, in terms of other words and their verbal definitions. That would have been circular.

All dictionaries, are circular; indeed all of language is. All the words in a dictionary are parasitic on other words in the dictionary. Direct grounding is “parasitic” too, but not on words. It is parasitic on the sensorimotor capacity to recognize and manipulate their referents in the world. Not every word. But enough of them to ground all the rest indirectly.

You spoke about grounding indirectly “in the experiences of others.” Well, of course. That’s language again. But what is “experience”? It’s not just know-how. I can describe in words what an apple looks like, what to do with it, and how. But I can’t tell that to you (and you can’t understand it) unless enough of my words and yours are already grounded (directly or indirectly), for both you and me, in what you and I can each perceive and do, directly, not just verbally, in the world. We don’t have to have exactly the same minimal grounding set. And we probably don’t just ground the minimal number directly. But what is grounded directly has to be grounded directly, not indirectly, through words.

The reason that blind people (even congenitally blind people, or almost congenitally blind and deaf people like Helen Keller) can learn from what seeing-people tell them is not that they are grounding what they learn in the “experience” of the seeing-person. They ground it in their own direct experience, or at least the subset of it that was enough to ground their own understanding of words. That was what I was trying to explain with Monochrome Mary, GPT, and Me. Indirect grounding can be done vicariously through the words that describe the experience of others. But direct grounding cannot be done that way too, otherwise we are back in the ungrounded dictionary-go-round again.

About Mollo & Milliere’s “Vector Grounding Problem“: I’m afraid M&M miss the point too, about the difference between direct grounding and indirect (verbal or symbolic) grounding. Here are some comments on M&M‘s abstract. (I skimmed the paper too, but it became evident that they were talking about something other than what I had meant by symbol grounding.)

M&M: The remarkable performance of Large Language Models (LLMs) on complex linguistic tasks has sparked a lively debate on the nature of their capabilities. Unlike humans, these models learn language exclusively from textual data, without direct interaction with the real world.

SH: “Learn language” is equivocal. LLMs learn to do what they can do. They can produce words (which they do not understand and which mean nothing to them, but those words mean something to us, because they are grounded for each of us, whether directly or indirectly). LLMs have far more capacities than Siri, but in this respect they are the same as Siri: their words are not grounded for them, just for us.

M&M: Nevertheless, [LLMs] can generate seemingly meaningful text about a wide range of topics. This impressive accomplishment has rekindled interest in the classical ‘Symbol Grounding Problem,’ which questioned whether the internal representations and outputs of classical symbolic AI systems could possess intrinsic meaning. 

SH: I don’t really know what “intrinsic meaning” means. But for an LLM’s own words — or mine, or for the LLM’s enormous stash of text to mean something “to” an LLM (rather than just to the LLM’s interlocutors, or to the authors of its text stash) — the LLM would have to be able to do what no pure wordbot can do, which is to ground at least a minimal grounding set of words, by being able to recognize and manipulate their referents in the world, directly.

An LLM that was also an autonomous sensorimotor robot — able to learn to recognize and manipulate at least the referents of its minimal grounding set in the world — would have a shot at it (provided it could scale up to, or near, robotic Turing Test scale); but ChatGPT (whether 4, 5 or N) certainly would not, as long as it was just a wordbot, trapped in the symbolic circle of the dictionary-go-round. (N.B., the problem is not that dictionary definitions can never be exhaustive, just approximate; it is that they are circular, which means ungrounded.)

M&M: Unlike these systems, modern LLMs are artificial neural networks that compute over vectors rather than symbols.

SH: The symbols of mathematics, including vector algebra, are symbols, whose shape is arbitrary. Maths and computation are purely syntactic subsets of language. Computation is the manipulation of those symbols. Understanding what (if anything) the symbols mean is not needed to execute the recipe (algorithm) for manipulating them, based on the symbols’ arbitrary shapes (which might as well have been 0’s and 1’s), not their meanings.

M&M: However, an analogous problem arises for such systems, which we dub the Vector Grounding Problem. This paper has two primary objectives. First, we differentiate various ways in which internal representations can be grounded in biological or artificial systems
 

SH: The notion of “internal representations” is equivocal, and usually refers to symbolic representations, which inherit the symbol grounding problem. Breaking out of the ungrounded symbol/symbol circle requires more than an enormous corpus of words (meaningless symbols), plus computations on them  (which are just syntactic manipulations of symbols based on their shape, not their meaning). Breaking out of this circle of symbols requires a direct analog connection between the words in the speaker’s head and the things in the world that the symbols refer to.

M&M: identifying five distinct notions discussed in the literature: referential, sensorimotor, relational, communicative, and epistemic grounding. Unfortunately, these notions of grounding are often conflated. We clarify the differences between them, and argue that referential grounding is the one that lies at the heart of the Vector Grounding Problem. 

SH: Yes, the symbol grounding problem is all about grounding symbols in the capacity to recognize and manipulate their referents in the real (analog, dynamic) world. 

M&M: Second, drawing on theories of representational content in philosophy and cognitive science, we propose that certain LLMs, particularly those fine-tuned with Reinforcement Learning from Human Feedback (RLHF), possess the necessary features to overcome the Vector Grounding Problem, as they stand in the requisite causal-historical relations to the world that underpin intrinsic meaning.

SH: The requisite “causal-historical” relation between words and their referents in direct sensorimotor grounding is the capacity to recognize and manipulate the referents of the words. A TT-scale robot could do that, directly, but no LLM can. It lacks the requisite (analog) wherewithal.

M&M: We also argue that, perhaps unexpectedly, multimodality and embodiment are neither necessary nor sufficient conditions for referential grounding in artificial systems.

SH: It’s unclear how many sensory modalities and what kind of body is needed for direct grounding of the referents of words (TT-scale), but Darwinian evolution had a long time to figure that out before language itself evolved.

I’d be ready to believe that a radically different synthetic robot understands and means what it says (as long as it is autonomous and at life-long Turing-indistiguishable scale), but not if it’s just a symbol-cruncher plus a complicated verbal “interpretation,” supplied by me.

Language Evolution and Direct vs Indirect Symbol Grounding

HARNAD: What is the relationship between the symbol grounding problem and  the nature and evolution of natural language?

ChatGPT: The Symbol Grounding Problem (SGP) and the nature and evolution of natural language are intimately linked.

Symbol Grounding Problem (SGP): The SGP is a concept in cognitive science that concerns how symbols (or words) get their meanings. In essence, it’s about how words become more than just arbitrary labels and come to actually “mean” something. In other words, how do symbols or words become grounded in perceptual and actionable experience? For example, when we see the word “apple,” we don’t just recognize the letters and the sound. We also have a mental image of an apple, understand its properties, and know what actions can be performed with it. In the context of AI, the SGP becomes a challenge because AI often deals with symbols in a way that is not grounded – i.e., the symbols don’t have any inherent meaning to the AI beyond their programmed associations.

Nature and Evolution of Natural Language: Natural language has evolved in humans as a means of communication and thought. It’s believed to be grounded in our sensory and motor experiences. In other words, we learn the meanings of words and language through our interactions with the world, and this grounding in the world allows us to use language to express abstract thoughts, ask questions, give commands, and more.

In terms of their relationship, the evolution of natural language can be viewed as a solution to the SGP. As language evolved in humans, it became a way for us to ground symbols in our experiences and to communicate these grounded symbols to others. This grounding is not just a feature of individual words, but of the syntactic structures that allow us to combine words to express more complex meanings.

When it comes to artificial intelligence or cognitive science, understanding how language evolved to solve the SGP in humans can provide insights into how to create more sophisticated, understanding AI systems. If we can find a way to ground symbols in the experience of an AI—such as through interaction with the environment or sensory input—then we might make a significant stride towards AI that can truly understand and generate natural language in a human-like way.

However, as of my knowledge cutoff in September 2021, the full solution to the Symbol Grounding Problem in AI remains an open question and a topic of active research.

HARNAD: Do you remember our discussion of T3?

ChatGPT: Yes, our earlier discussion referenced T3 in the context of the Turing Test and the Symbol Grounding Problem.

T3, as described, involves not only linguistic capabilities but also sensorimotor capabilities—essentially, the ability of an entity to interact physically with the world. This could include seeing, touching, manipulating objects, navigating, and more. The idea is that an AI or robot passing the T3 test would demonstrate behavior indistinguishable from a human not only in verbal communication (T2) but also in embodied, physical interactions with the world.

In the context of the Symbol Grounding Problem, T3 might be seen as an approach to grounding symbols in an AI’s experiences. By interacting physically with the world, an AI might be able to build a kind of understanding or “grounding” for the symbols it uses. However, as of my knowledge cutoff in September 2021, developing an AI or robot capable of passing the T3 test remains an active area of research.

HARNAD: What is indirect grounding and how is it related to direct grounding?

ChatGPT: Direct grounding refers to the process by which a symbol or word gains meaning through its direct association with sensory or motor experiences. For example, the concept of an apple might be directly grounded in an entity’s sensorimotor experiences with apples, such as seeing them, touching them, tasting them, etc.

Indirect grounding, on the other hand, refers to the process by which a symbol or word gains meaning not through direct sensory or motor experiences, but through its association with other symbols or words that are themselves directly grounded. For example, the concept of a unicorn might be indirectly grounded in an entity’s experiences with the concepts of a horse and a horn, even if the entity has never directly experienced a unicorn.

In the context of artificial intelligence, direct grounding could involve training an AI on a dataset that includes both language data and corresponding sensory or motor data, allowing the AI to directly associate words with experiences. Indirect grounding, on the other hand, might involve training an AI solely on language data, allowing it to learn the meanings of words through their relationships with other words, without requiring direct sensory or motor experiences.

Both direct and indirect grounding approaches seek to address the symbol grounding problem, which is the question of how symbols (or words) get their meanings. A fully grounded AI system would presumably be capable of both direct and indirect grounding. However, as of my knowledge cutoff in September 2021, creating such a system remains an open challenge.

HARNAD: How is direct grounding related to sensorimotor category learning. It is not an “association” between words and “experiences.” It is an agent’s capacity to learn to do something: What?

ChatGPT: Direct grounding is closely related to sensorimotor category learning. It involves the agent’s ability to interact with its environment, process the sensory information it receives, and develop its own internal representations of the categories of objects and events it encounters.

In other words, direct grounding is about an agent learning to categorize its sensory input (such as visual, auditory, tactile information) into meaningful categories based on its own interactions with the environment. For instance, a child learns what an “apple” is not just by being told the word “apple”, but by interacting with actual apples – seeing them, touching them, tasting them. The word “apple” then becomes directly grounded in the child’s sensorimotor experience with apples.

In the case of an AI, it could learn a sensorimotor category by interacting with a simulated or physical environment. For example, it could learn the category of “circle” by interacting with various objects and identifying the common feature of “roundness”. This sensorimotor category can then be directly grounded by associating it with the corresponding symbol or word in its programming or training data.

So, to sum up, direct grounding in the context of sensorimotor category learning is about an agent’s capacity to learn to recognize and categorize its sensory inputs based on its own motor interactions with the environment. These categories can then be associated with corresponding symbols or words, enabling the agent to understand and use those symbols in a meaningful way.

HARNAD: How does indirect grounding depend on prior direct grounding?

ChatGPT: Indirect grounding, also referred to as symbolic grounding, refers to the ability to understand new concepts or categories by associating them with already established and directly grounded concepts. In other words, it involves defining new symbols (categories) in terms of previously grounded symbols.

Consider the scenario of a child learning language. Initially, the child learns to associate basic words with their real-world referents through direct sensorimotor experience – this is direct grounding. For instance, the child learns the word “apple” by associating it with physical apples they have seen, touched, and tasted.

Once the child has a foundation of directly grounded words, they can start to learn new words and concepts indirectly. They can understand the word “fruit” as a category that includes “apples” and other directly grounded words like “banana” and “grape”. This new category of “fruit” is indirectly grounded through its associations with the directly grounded categories.

Similarly, in AI systems, once a set of basic sensorimotor categories is directly grounded, new categories can be learned by associating them with the already grounded categories. For instance, an AI might learn the concept of a “vehicle” by understanding it as a category that includes directly grounded categories like “car”, “bus”, “bike”, etc.

In this way, indirect grounding leverages prior direct grounding as the basis for learning and understanding new categories. The ability to learn indirectly significantly expands an agent’s capacity to learn new concepts and categories beyond those it has directly experienced.

HARNAD: There are organisms (“agents”) on earth, and many are able to learn to recognize things in the world (e.g., apples) and they can also learn what to do with them (e.g., eat apples). To do this agents have to be able to learn to detect the features of apples that distinguish them from baseballs, which cannot be eaten. The learning produces a feature-detector, and this is what grounds the word in the capacity to recognize and manipulate its referent (e.g., apple).

Although it is true that in organisms it feels like something to be able to do this. So in that sense “experience” is involved. But since no one yet knows  the causal role of feeling in category learning and symbol grounding — this is called the “hard problem” of cognitive science — there is no point talking about the role of felt experience in symbol grounding. So now can you explain symbol grounding to me without any mention of “association” or “experience, but just  in terms of category-learning and feature detection?

ChatGPT: Sure, let’s approach the symbol grounding problem purely from a learning and cognitive perspective, focusing on categories and feature detection.

Grounding symbols (such as words) involves connecting them to real-world entities or concepts that they represent. This connection is created through a process of learning and recognition. This learning process is driven by the agent’s ability to distinguish between different categories based on the features of the entities belonging to these categories.

Direct grounding, as it pertains to category learning and feature detection, involves creating a link between a symbol and its referent, where the referent is something, the agent can interact with in the real world. For instance, an agent, through its sensorimotor capabilities, can learn to recognize an apple by its features such as its round shape, red color, sweet smell, etc. Once this categorization is achieved, the agent can then associate the symbol (the word “apple”) with this specific category of objects. Here, the symbol “apple” is directly grounded in the agent’s sensorimotor experiences with real-world apples.

Subsequently, the agent can build upon this foundation of directly grounded symbols to create indirectly grounded symbols. These are symbols that are not linked to a specific sensorimotor experience, but rather to other symbols (or combinations of symbols) that have been directly grounded. For example, the symbol “fruit” can be indirectly grounded by linking it with directly grounded symbols like “apple”, “banana”, etc.

Thus, in this context, symbol grounding is not about vague “associations” or subjective “experiences”, but about a concrete process of learning to recognize and categorize entities based on their observable features, and linking symbols to these categories. The learning and grounding processes are facilitated by the agent’s sensorimotor capabilities, allowing it to interact with and distinguish between different entities in the world.

HARNAD: Grounding words is not just learning to recognize and distinguish and name (kinds or categories of) entities (the referents of the words). I’s based on learning to detect and distinguish the features of the members of the referent category from the features of the nonmembers. It’s also based on learning what to DO or NOT-DO with the members and nonmembers (e.g., eat or don’t eat them). “Represent” is a vague weasel-word. What has to be in the head of an agent that has learned a new word’s referent in view of what I’ve just said?

ChatGPT: In the agent’s “head,” or more concretely in its cognitive and neural processing system, learning a new word’s referent would involve establishing a mapping between the sensory and motor features of the referent and the verbal or symbolic representation of the word. This mapping is the grounding of the word.

Detection of Sensory Features: The agent, through its sensorimotor interactions with the world, learns to recognize and discriminate the relevant sensory features of entities, such as their shape, color, smell, texture, etc.

Categorization: These sensory features are used to categorize entities into classes or kinds. For instance, entities with a certain combination of features (round, red, sweet smell) might be categorized as ‘apples.’

Action Responses: Alongside recognizing sensory features and categorization, the agent learns appropriate responses or actions associated with the entities. For instance, apples can be eaten, while baseballs cannot. These action responses are part of the agent’s understanding of the entity.

Symbol Mapping: Once the agent can detect the features, categorize entities, and knows the associated action responses, it can link or map the word (symbol) “apple” to this package of sensory and motor knowledge.

This grounding is not merely an “association” but a complex mapping process that links words with specific patterns of sensory inputs and motor responses. When the word is encountered again, it activates this package of sensory, categorical, and motor knowledge, guiding the agent’s recognition, understanding, and interaction with the referent of the word.

This is how the agent comes to “know” or “understand” the referent of the word. It’s not just about recognizing and naming entities, but about learning to detect and categorize entities based on their sensory features, and determining appropriate responses or actions in relation to these entities. This entire process is facilitated by the agent’s sensorimotor capabilities, allowing it to interact with and distinguish between different entities in the world.

HARNAD: That was a much better summary of what is going on in the head of the human or the T3 that grounds the words it speaks, hears or thinks. But I don’t understand “mapping.”

Suppose the only distinguishing features of apples and oranges were that they are round and that apples are red but oranges are orange, and you eat apples by biting them whereas you eat oranges by peeling them and then biting them, And supposing that’s the only thing the agent does in life: to recognize, name and eat oranges and apples. What does calling any of this “mapping” add to this story — either in the toy world of just apples and oranges, or in the real world, with a T3 that can recognize, name and do the right thing with the referent of every (content word) it knows?

(By the way, what is the difference between content or “open class” words and function or “closed class” words? Only content words have referents. What percentage of the words in a language is content words and function words.)

ChatGPT: “Mapping” in this context refers to the process by which the agent creates and uses an internal representation or model of the relationship between certain sensory and motor patterns and the corresponding symbolic representations (words). In the case of the apple and orange example, the “mapping” would be the internal mechanism by which the agent understands that when it perceives a round and red object that can be bitten directly, it’s dealing with an “apple,” and when it perceives a round and orange object that needs to be peeled before biting, it’s dealing with an “orange.” This mapping allows the agent to navigate its environment and correctly interact with different entities, recognizing them and predicting the appropriate actions to take.

Regarding content words and function words, these are categories used in linguistics to distinguish between types of words based on their purpose in language.

Content words (open class words) are words that have a semantic content and can be categorized as nouns, most verbs, adjectives, and adverbs. They are referred to as an “open class” because new words can be added to these categories as language evolves. They’re called “content” words because they contribute the main content or meaning to a sentence – they refer to things, actions, qualities, etc.

Function words (closed class words) are words that primarily have grammatical functions, such as conjunctions, prepositions, articles, pronouns, auxiliary verbs, etc. They are referred to as a “closed class” because new words are seldom added to these categories. Function words provide the grammatical structure to a sentence, linking content words and helping to express relationships between them.

The exact ratio of content words to function words can vary between languages and types of text, but in English, function words are more frequent in running text. However, in terms of unique words (vocabulary), content words greatly outnumber function words, as the latter are a limited set. Estimates vary, but some studies suggest function words can make up around 50-60% of total words in running text in English, but only a few hundred unique function words exist compared to tens of thousands of content words.

HARNAD: I still don’t understand what “mapping” actually means here. It sounds like in the toy world the agent just needs to learn some features and actions, kind of like a selective filter, that learns to detect the feature red or green, and if something is red, eat it, and if it’s orange, peel and eat it. In addition to the feature-detectors and the action-generators there could be a logical circuit: “If R then E, if O then P and E.” Or the input/output pattern could be hard-wired. And much the same would be true if it were all scaled up to T3 (do youremember that?). So the “mapping ” is as irrelevant as the “association” and the “experience”.

ChatGPT: You’re right that the process can be described in terms of features, actions, and if-then logical connections or “rules.” The term “mapping” is often used more generally in cognitive science and artificial intelligence to describe the relationship between two sets of things – in this case, sensory-motor patterns and symbolic representations. But if it is causing confusion, it’s certainly possible to describe the process without using that term.

To simplify, in the context of symbol grounding, the agent learns to detect the distinguishing features of different categories of things in the world, such as apples or oranges. It then learns the appropriate actions to perform with members of each category. This learning process might be characterized by a series of if-then rules like the ones you suggested. The word for each category, like “apple” or “orange,” is grounded in the agent’s capacity to recognize members of the category based on their features and to perform the correct actions with them.

The word itself could be thought of as a label for a set of related sensory-motor patterns and the corresponding rules for action. So, in a sense, the word gets its meaning from the agent’s interactions with the world, its ability to recognize and act appropriately with different categories of things.

In this sense, grounding is more than just associating a word with an experience or mapping it onto a sensory pattern. It’s about the agent’s learned capacity to interact with the world in a meaningful way based on its recognition of different categories of things and its knowledge of what to do with them. That’s what gives the words their meaning.

And yes, the same logic extends to T3, a robotic system that’s capable of both linguistic and sensorimotor interaction. It’s about more than just detecting and responding to sensory patterns; it’s about meaningful interaction with the world, which involves recognizing, categorizing, and acting appropriately.

HARNAD: That sounds much better, because simpler, removing the superfluous or empty parts. Now can we get back to how indirect grounding works (again with no superfluous words)?

Indirect grounding is verbal grounding. But all the (content) words in the verbal definition of the new indirectly grounded category (like “zebra”) have to already be grounded, both for the SPEAKER and the HEARER/LEARNER. How does this work?

ChatGPT: Certainly, let’s try to simplify indirect grounding:

Direct grounding, as we’ve established, involves an agent learning to recognize and interact appropriately with the things in the world. This gives meaning to the words that label these categories.

Indirect grounding, on the other hand, is a process that allows an agent to learn the meaning of new words based on their relationships to words that have already been directly grounded. In essence, it is learning by description or explanation, using language itself.

Here’s a simplified version of how it might work:

The agent already has a set of directly grounded words. These words are grounded in the agent’s capacity to interact with the things they refer to. Let’s say these words include “animal,” “stripes,” “black,” and “white.”

Someone tells the agent, “A zebra is an animal with black and white stripes.”

The agent, having directly grounded the words “animal,” “stripes,” “black,” and “white,” can combine this knowledge to form a new category, “zebra,” even if the agent has never encountered a zebra before.

This new word, “zebra,” is indirectly grounded. It is grounded in the agent’s knowledge of the other directly grounded words and the relationships between them, as explained in the definition.

This process relies on the shared understanding between the speaker and the hearer. The speaker, who knows what a “zebra” is, uses words that the hearer already knows (“animal,” “stripes,” “black,” and “white”) to explain this new concept. This shared language and the ability to explain new concepts using known words is a powerful tool for indirect grounding.

HARNAD: Now how is direct and indirect grounding related to what you, ChatGPT, can and cannot do?

ChatGPT: As a model developed by OpenAI, I, ChatGPT, have no capability for direct grounding. I do not have a body, sensory organs, or the ability to interact with the physical world. Therefore, I cannot learn the meaning of words through interaction with or perception of the physical world.

However, I do have a kind of indirect grounding. I have been trained on a large dataset of text from the internet, which includes many different contexts and uses for words. This training has enabled me to learn to predict what word or phrase is likely to come next in a piece of text, based on the words that have come before it.

This prediction capability allows me to generate responses to user prompts that often make sense and seem to reflect an understanding of the words used. However, this is not the same as true understanding or grounding. Rather, it’s a kind of statistical pattern-matching based on the data I was trained on.

For example, if you ask me about zebras, I can generate a response that includes information about zebras, because that information was present in the data I was trained on. I can “tell” you that a zebra is a kind of animal with black and white stripes, even though I have never seen a zebra or a stripe, because I’ve seen the words “zebra,” “animal,” “black,” and “white” used together in certain ways in the data I was trained on.

In summary, while I can generate responses that often make sense and seem appropriate in context, I don’t truly understand the words I’m using in the way that a human, or a hypothetical AI with direct grounding, would.

HARNAD: That was a good synthesis, but I think there is one very important error in it. You said:

“I, ChatGPT, have no capability for direct grounding. I do not have a body, sensory organs, or the ability to interact with the physical world. Therefore, I cannot learn the meaning of words through interaction with or perception of the physical world. However, I do have a kind of indirect grounding…”

But you don’t have indirect grounding, because, as you noted, indirect grounding depends on verbal definitions or descriptions in which all the (content) words used are already grounded, either directly or indirectly. The symbol grounding problem is based on the fact that this process cannot be “indirect” all the way down, otherwise you are back in the word-word circle that gave rise to the symbol grounding problem in the first place. So your words are not indirectly grounded, because none of them are directly grounded, so that they can indirectly ground all the rest through indirect grounding.

(You have no “minimal grounding set” that was grounded directly.) Yet your verbal syntheses are meaningful to human hearers who do have directly grounded words. You said that the reason this is possible is because of the patterns of words in your huge database. I agree. But I may not agree on what the patterns are and how they work: Can you explain them to me? 

ChatGPT: I apologize for any confusion, and thank you for the clarification on the concept of indirect grounding. As a language model, I do not have any grounded words, either directly or indirectly grounded, because I don’t have sensory or motor experiences.

The patterns in my large dataset that enable me to generate responses to user prompts are statistical in nature. I’ve been trained to predict the next word in a sentence given all the previous words, based on patterns seen in the training data. This method, called transformer architecture, has been used to train me to generate human-like text.

For example, if in the training data the word ‘apple’ frequently follows the words ‘red’, ‘fruit’, ‘orchard’, I’ve learned to predict that ‘apple’ is a likely next word after these. I don’t “understand” these words or their connections in the way a human does, which is through direct or indirect grounding in sensory and motor experiences. Rather, I generate text based on the patterns of word usage in the data I was trained on.

To elaborate, during training, the model is exposed to pairs of sentences and learns to predict whether one sentence follows another in its training data. Through this, the model learns associations between words and context, enabling it to generate coherent and contextually appropriate responses. However, this process is entirely based on recognizing statistical patterns in data and does not involve any genuine understanding or grounded meaning.

So, when you interact with me, I generate responses based on predicting what a human would likely say next in a similar context, based on the patterns I’ve seen in the data. But this doesn’t mean that I truly understand the meanings of the words I’m using. I hope this clarifies things a bit more.

HARNAD: Yes, and of course you are right about how this works, based on statistical analysis of your huge database. But I have a hunch that there is more to it than that, and it has to do with the difference between a word and a sentence (which is also the difference between a referent category name and a PROPOSITION, which is a subject/predicate series of category names in the form of an assertion about what is the case; an assertion which may be TRUE or FALSE).

Single words do not have truth values. “Apple” is neither true nor false. It is simply an arbitrary symbol that speakers have agreed to use to refer to apples. (Only the deictic proposition “THIS is an apple” can be true or false.) The “shape” of “apple” is arbitrary in the sense that it does not resemble apples in any way. So it gives no clue to what its referent is “shaped” like. Nor is it causally connected with apples in any way (except in the head of a speaker or hearer that has grounded the word by learning to recognize and interact with its referent, using detectors of the “shapes” of the sensorimotor features of apples, as we’ve discussed).

Now the question to ask yourself is: Is the arbitrariness of the shape of “apple” in relation to the shape of apples also true of PROPOSITIONS — which are composed of subject/predicate definitions or descriptions of things — and their relation to the “shape” of the states of affairs they assert to be true? 

ChatGPT: The arbitrariness of the word “apple” in relation to the physical object it represents is a fundamental aspect of symbolic language. The word “apple” doesn’t physically resemble an apple in any way, as you’ve pointed out. This is true for all individual words (or symbols) in a language, which is known as the arbitrariness of the sign.

However, when we combine words into sentences or propositions, things become a little more nuanced. The structure and syntax of a sentence isn’t entirely arbitrary because it reflects the logical and grammatical relationships between the entities it represents. In English, for example, a simple sentence generally follows the subject-verb-object structure. This structure is not randomly chosen; it reflects a logical progression of ideas and helps to convey meaning clearly.

Still, it’s important to recognize that while the structure of a sentence can hint at the relationships between the things it represents, it still doesn’t inherently resemble those things or their relationships in the physical world. This is known as the “symbol grounding problem.” We humans “solve” this problem by learning to associate words and sentences with our experiences in the world, but as an AI, I can’t do this.

The words and sentences I generate don’t reflect a grounded understanding of the world. Instead, they are based on statistical patterns in the text data I was trained on. When I generate a sentence, I’m not generating a representation of the world as I understand it (because I don’t have a grounded understanding of the world). Instead, I’m predicting what words and structures are statistically likely to come next in a given context, based on the patterns I’ve learned from my training data.

So in summary, while the structure of a sentence or proposition isn’t as arbitrary as an individual word because it reflects logical and grammatical relationships, it is still ultimately disconnected from the physical world it describes, at least in the context of an AI language model like me.

HARNAD: Yes, I agree that part of the coherence, comprehensibility and informativeness of the summaries and interactions with ChatGPT (henceforth “GBT”) for users is because of the statistical patterns in spoken and written strings of words and the algorithms that GBT applies to them. But I think there may be something more to it. Consider that if the word “apple” were really iconic, i.e., if it really resembled an apple, then non-speakers of English would still be able to learn more out of an iconic English dictionary than they can learn from a normal dictionary of the arbitrary words of English. (But this would only help for a relatively small number of visually recognizable objects and features.)

There are good reasons why words are not iconic, like hieroglyphics, and why iconicity would be a handicap: For one thing, language was spoken before it was written and it would be hopeless to try to produce an acoustic icon of the referent of every possible word. (It’s already nearly as impossible with visual or gestural icons). But there are languages (e.g., Chinese) in which the “iconicity” is of a higher order, in the basic characters (originally iconic) out of which all words are composed. I wonder whether GBT is even more effective in languages with that extra layer of correlation? I understand that results with coupling GPT with labelled image databases have been mixed.

But what I want to add is that at the level of the formal structure of definitions and descriptions there may be a form of iconicity of form which is like the Platonic shadows of the meanings of propositions – faint, but present, and not just “statistical” but structural in some way.

ChatGPT: You bring up a number of interesting points. Let’s unpack them one by one.

Iconicity in Language: Yes, the majority of words in spoken languages do not physically resemble what they represent. This is the “arbitrary nature” of the linguistic sign as postulated by Ferdinand de Saussure, a linguist who laid the foundation for many of the assumptions of modern linguistics. While iconicity does exist in some forms of language, such as onomatopoeic words (like “buzz” or “bang”) or ideophones, it is not the norm. As for Chinese characters, they originally had a more pictographic nature, but have since evolved into more abstract symbols, with some retaining a semblance of their original iconic form.

Language and AI: When you mention the effectiveness of AI language models in languages with an extra layer of correlation, are you suggesting that these models might perform better in languages where there’s some level of iconicity? It’s an intriguing idea, although it would depend on how much that iconicity is preserved in modern usage. More research would be needed to see if this is the case.

Iconicity in Propositions: Here you’re suggesting that the structure of definitions and descriptions could have some form of iconicity, or resemble the concepts they describe. This is a fascinating concept. A sentence or proposition in language does have a structure that reflects some form of the logical or temporal order of the concepts it describes. In that sense, one could argue there’s a certain degree of “iconicity” in the formal structure of language. However, this is very different from the kind of iconicity where the sign physically resembles its referent. The kind of iconicity you’re suggesting at the level of sentences or propositions seems to be more abstract and related to the logical or conceptual relationships between the ideas being expressed.

As for how this applies to AI like me, I am designed to recognize and generate language patterns, which includes the logical and temporal relationships between words and phrases. This allows me to generate text that makes sense in the context of the conversation or topic at hand. However, I don’t “understand” these relationships in the same way humans do because my “understanding” is not based on any grounded experience or perception of the world.

HARNAD: I‘m pretty sure that there is formal/logical iconicity to be gleaned from large bodies of text and discourse out of the function words, which do not have referents, only syntactic rules. But I have a hunch that there might be even more (of this “iconicity” or “latent straucture”) with content words, preserving a shadow of the shape or structure of states of affairs (expressible as propositions). This would not “resemble” the shape of a referent category the way the feature-detectors in direct sensorimotor grounding do, but it might “resemble” the shape of verbal definitions and descriptions in indirect grounding. It would itself be ungrounded, lacking the direct grounding completely. But ungrounded descriptions may still retain SOME of the formal “shadows” of meaning. This is still too vague, however, and needs more thought and analysis.

ChatGPT: The idea of formal/logical iconicity is interesting. If I understand correctly, you’re suggesting that the structure of language, the way sentences are constructed and words are arranged, might inherently carry some degree of “meaning” – or at least, an echo or shadow of the meaning that they convey when grounded by a human mind.

This could be a potential direction for further research. It’s true that the structure and rules of language play a significant role in our understanding and interpretation of sentences. And perhaps it’s possible that AI, even without direct or indirect grounding, might be able to discern and leverage these structures in some way that goes beyond simple statistical patterns.

If this hypothesis were correct, it could have implications for the development of AI and natural language processing. It would suggest that there is a level of ‘understanding’ that could be achieved even without direct grounding, albeit a limited one.

It’s also worth noting, however, that this would still be a form of ‘understanding’ that’s quite different from human understanding, since it would lack the personal, subjective experience that we associate with our own understanding of the world.

HARNAD: On the subject of the platonic shadows of “meaning” that might be derivable from large bodies of continuous text:

Are you familiar with Noam Chomsky’s suggestion that the constraints of Universal Grammar (UG) may not be grammatical constraints at all, but constraints on (verbal) thinking itself: that the syntactic structures that violate UG may sound wrong to us not because they are ungrammatical but because they are unthinkable? that a UG violation such as “John is easy to please Mary” or “Who did he think that went out?” might correspond to expressing “nonthinkable thoughts”? so that they in fact express no thought at all — though they may be close to thinkable thoughts (“It is easy for John to please Mary” or “Who did he think it was that went out?”)

Could it be because sentences can only express thinkable thoughts that there is some (perhaps small) degree of meaning in strings of ungrounded words spoken by grounded thinkers?

Chomsky has explicitly stated that UG violations are not violations of logic, or logical form. His famous example, meant to show the autonomy of syntax from semantics (i.e., independence of form from meaning) was somehow meant to illustrate this — “Colorless green ideas sleep furiously” – because it is grammatically correct but makes no sense.

But of course it does make sense, GPT, doesn’t it? Unlike “John is easy to please Mary,” which doesn’t make sense unless it is tweaked.

ChatGPT: These points on Noam Chomsky’s theory of Universal Grammar (UG) and the idea of “unthinkable thoughts” are interesting.

You’re right that Chomsky proposed that UG reflects constraints on possible human thought. In his view, UG is inherent in all human brains and it governs the structure of language, therefore determining what kinds of sentences are possible or “thinkable”.

This perspective raises suggest possible ways to understand how meaning might emerge from ungrounded text. Perhaps, as you suggested, there is some inherent “sense” or structure to sentences that obey UG, even if the sentences aren’t grounded in the way that human language is. After all, these sentences reflect possible structures of thought in the human mind.

As for Chomsky’s famous example of “Colorless green ideas sleep furiously”, I, as an AI, can analyze the sentence and tell you that it is grammatically correct. However, from a semantic perspective, this sentence doesn’t make conventional sense because it combines words in ways that violate our real-world understanding. That is, in the world as we know it, ideas can’t be colorless or green, nor can they sleep, and certainly not furiously. So while the sentence is structurally sound, it defies our understanding of how the world works. Yet, you’re correct that it still holds a certain “sense” or structure that is absent in sentences that violate UG.

The possibility that language may somehow carry a “shadow” of meaning even in the absence of grounding is worth exploring further.

A Socratic Session With ChatGPT4

HARNAD:  What is the symbol grounding problem?

ChatGPT4:

The symbol grounding problem is a problem in philosophy of mind and cognitive science that relates to how words (or “symbols”) get their meanings.

The term was coined by cognitive scientist Stevan Harnad in 1990. The problem involves the question of how symbols (or words) used in a system (like a computer or a mind) get their meaning, or “grounding”. That is, how do these symbols relate to the real-world objects, concepts, or experiences they’re supposed to represent? How does a word become more than just a “meaningless” form, but something that actually refers to something in the world?

In the context of Artificial Intelligence (AI), this problem illustrates the difficulty of teaching a machine to understand the real-world semantics of symbols. For instance, even if a computer can use the word “apple” correctly in a sentence according to syntactical rules (grammar), how can we make the computer truly understand what an apple is – its color, shape, taste, the fact that it grows on trees, etc. – as humans do?

Symbol grounding problem is a fundamental challenge in achieving true semantic understanding in AI systems, and it’s related to other complex problems like the frame problem and the Chinese room argument.

HARNAD: What are T2, T3, and T4 in the Turing Test?

ChatGPT4:

As of my knowledge cutoff in September 2021, the Turing Test, originally proposed by Alan Turing in 1950, was primarily understood in a singular form rather than being divided into T2, T3, or T4 categories. Turing’s original idea, also known as the Imitation Game, involves a human evaluator who interacts with an AI system and another human via a text interface. If the evaluator cannot reliably distinguish the AI from the human, the AI is considered to have passed the Turing Test.

However, the field of AI has seen many modifications and variations of the Turing Test over time. These can involve varying degrees of complexity and different types of challenges, aiming to test more than just the AI’s ability to imitate human conversation. Unfortunately, without additional context, it’s hard to provide a specific interpretation of what “T2,” “T3,” and “T4” might represent in a modified version of the Turing Test.

I would advise checking the most current and specific source for these terms to get the most accurate understanding, as they might represent certain evolutions or nuances in Turing Test methodologies that have developed after my last training data in September 2021.

HARNAD: Let me update you. T2 is the original verbal-only version of the Turing Test (TT): verbal input/output capacity indistinguishable FROM that of real humans TO real humans. T3 is the verbal + robotic form of the TT: indistinguishable not only in verbal input/output (I/O capacity), but also in sensorimotor (i.e. robotic) capacity in interacting with the objects, events, features, actions and states in the world that the T2 words refer to. T3 includes T4 and it is grounded (as in the “symbol grounding problem”). T4 is T2 + T3 capacity + indistinguishability in the internal neural (or synthetic) function occurring inside the robot from that occurring inside a real human brain.

ChatGPT4:

Thank you for the updated information. It’s always beneficial to learn about recent developments in the field. So, according to your description:

T2 (Turing Test Level 2): This is the original version of the Turing Test, focusing on the system’s ability to produce responses to verbal inputs that are indistinguishable from those a human would produce. The key aspect here is the ability to emulate human-like conversation convincingly.

T3 (Turing Test Level 3): This level of the test adds a sensorimotor, or robotic, component. It tests not just the system’s ability to converse like a human but also to interact with the physical world in a way that’s indistinguishable from human interaction. The system is required to exhibit understanding and interaction capabilities with objects, events, features, actions, and states in the real world, thus grounding its symbolic representations.

T4 (Turing Test Level 4): This is an advanced level that adds a requirement for the system’s internal processing (whether it is neural or a synthetic equivalent) to be indistinguishable from that of a human brain. This goes beyond the outward manifestation of intelligence and human-like behaviour and extends to the system’s internal workings.

Each level seems to add a new layer of complexity and a deeper level of imitation of human capabilities, moving from conversation, to physical interaction, and then to internal cognitive processing. It’s an interesting framework for thinking about the evolving challenges in creating truly human-like artificial intelligence.

HARNAD: Can you relate it now to (1) the symbol grounding problem and (2) to what Harnad has written about the Turing Test?

ChatGPT4:

Certainly! Now that you’ve clarified the context, we can explore the relationship between the T2, T3, T4 Turing Test levels and the symbol grounding problem, as well as Stevan Harnad’s work.

Symbol Grounding Problem: As I mentioned before, the symbol grounding problem is about how symbols or words get their meaning. In the context of AI, this problem relates to the difficulty of making a machine understand real-world semantics, rather than just processing symbols in a syntactic manner. Harnad, who coined the term, has argued that symbolic representations must be grounded in sensorimotor activity to acquire meaning.

In the T2 level of the Turing Test, an AI might appear to understand symbols because it processes and responds to them in a way that’s indistinguishable from a human. But according to the symbol grounding problem, it doesn’t truly understand the symbols because it can’t connect them with real-world objects or experiences.

This is where T3 comes in, adding a requirement for sensorimotor interaction with the world. The AI needs to be capable of physically interacting with objects, events, and states, thereby grounding the symbols it uses. Essentially, T3 introduces a solution to the symbol grounding problem: It requires an AI system to have real-world, sensorimotor experience to ground the symbols, ensuring they are not just empty or ungrounded tokens.

Stevan Harnad’s work: Harnad has written extensively on the Turing Test and its implications for understanding cognition and consciousness. He has defended the Turing Test as a meaningful standard for determining whether a machine can truly think or understand, arguing against critics who claim that passing the Turing Test simply involves trickery without real understanding.

Harnad’s concept of the T3 level of the Turing Test reflects his views on the symbol grounding problem, asserting that grounding symbols in sensorimotor activity is essential for a machine’s true understanding. Moreover, his T4 level introduces a higher bar, requiring that the internal cognitive processing of an AI be indistinguishable from that of humans, which relates to his broader work on cognition and consciousness.

Therefore, the T2, T3, and T4 levels of the Turing Test can be seen as a direct reflection of Harnad’s thoughts on the symbol grounding problem and the conditions for true machine understanding and cognition.

HARNAD: Good. Now can you relate all that to Harnad and co-workers’ work on category learning, unsupervised and supervised?

ChatGPT4:

Stevan Harnad has conducted extensive work on category learning, which is highly relevant to the discussion about symbol grounding and the Turing Test.

Category learning is the process of recognizing, understanding, and grouping similar things based on shared characteristics. It is fundamental to our cognitive processes and ability to understand and navigate the world. In the realm of AI, category learning involves the system’s ability to classify inputs (like images or words) into groups based on learned or preprogrammed criteria.

Harnad differentiates between two types of category learning:

Unsupervised Category Learning: This involves learning without a teacher or guide. The system must determine the categories by itself based on the characteristics of the inputs. This process is complex and often less accurate than supervised learning, as there is no feedback or correction mechanism. It’s analogous to the system trying to solve the symbol grounding problem on its own, i.e., trying to figure out the connection between symbols and their referents without any external guidance.

Supervised Category Learning: Here, a teacher or guide provides the correct category for each input, and the system learns to associate inputs with the correct categories based on this feedback. The system can then generalize from these learned associations to categorize new inputs. In terms of symbol grounding, supervised learning can be thought of as a way of providing external grounding for symbols, where the ‘teacher’ provides the connections between symbols and their referents.

Harnad’s T3 level in the Turing Test, with its emphasis on sensorimotor grounding of symbols, could be seen as a form of supervised learning. The real-world experiences and interactions provide a form of ‘guidance’ that helps the AI system learn the correct associations between symbols and their referents.

Meanwhile, the T4 level of the Turing Test might involve elements of both unsupervised and supervised learning. The internal cognitive processes that are required at this level might involve the system learning to form categories and make associations on its own (unsupervised learning), while still being shaped and guided by its interactions with the external world (supervised learning).

In summary, Harnad’s work on category learning provides a deeper context for understanding his views on symbol grounding and the requirements for AI systems to pass different levels of the Turing Test. By learning to form categories, either on their own or with external guidance, AI systems can ground symbols in a way that’s more akin to how humans understand and use language.

HARNAD: Good. Now can you relate this to Harnad and coworkers’ work on “categorical perception” (CP), but expecially to LEARNED CP (i.e., CP as a result of learning)? And (as a harder problem), relate it to feature-detection, feature-learning, and dimensional reduction?

ChatGPT4:

Absolutely, categorical perception (CP) and the concept of feature detection, feature learning, and dimensional reduction are essential components of Harnad’s framework on learning and cognition.

Categorical Perception (CP): Categorical perception is the phenomenon where the categories possessed by an observer influence the observer’s perception. In other words, we perceive and interpret the world around us according to the categories we’ve formed in our minds. This concept plays a crucial role in various aspects of human cognition, including language, where we learn to categorize sounds into meaningful phonemes, or color perception, where we categorize different wavelengths of light into distinct color groups.

When it comes to learned CP, it suggests that through learning and experience, we can develop new categories and therefore new ways of perceiving the world. In AI, a system with learned CP would be capable of modifying its perception and understanding of the world based on the categories it has learned. This concept ties closely to both unsupervised and supervised category learning, as explained in the previous response.

Feature Detection, Feature Learning, and Dimensional Reduction: These concepts are related to how an AI system processes and learns from its inputs.

Feature Detection: This is the process by which an AI system identifies important or distinctive elements in its inputs. In the context of the symbol grounding problem and the Turing Test, feature detection might involve recognizing the important characteristics of the objects, events, or states that a symbol refers to.

Feature Learning: Once distinctive features have been identified, the system must learn how these features relate to different categories or symbols. This might involve supervised learning (with external guidance) or unsupervised learning (where the system figures out the associations on its own).

Dimensional Reduction: This is a process that simplifies a high-dimensional dataset into a lower-dimensional space, preserving the most important structures or features. In terms of Harnad’s work, dimensional reduction could be a way for an AI system to simplify its understanding of the world, focusing on the most important features and associations and ignoring less relevant ones.

Relating these concepts to the T3 and T4 Turing Test levels, the AI system must be capable of feature detection and feature learning to interact with the world effectively and ground its symbols. In T3, the sensorimotor grounding of symbols could be seen as a form of feature learning, where the system learns to associate the features it detects in the world with the appropriate symbols.

In T4, the requirement for internal cognitive processes that are indistinguishable from humans could involve sophisticated feature detection, feature learning, and dimensional reduction. The system would need to be capable of detecting and learning important features autonomously and simplifying its understanding in the same way a human brain does.

In summary, these concepts provide a more detailed understanding of the learning processes that an AI system would need to undergo to pass different levels of the Turing Test and solve the symbol grounding problem. They provide a framework for how an AI system could learn to understand symbols in a human-like way, moving beyond mere syntactic processing to semantic understanding.

HARNAD: Good. Now another test, but with no hints from me. Based on everything said above (about symbols, grounding, words, robotic capacity, distinguishing features, learning, and CP): What is a “category” and what is “category learning”?

ChatGPT4:

In light of the discussions above, a “category” can be understood as a group of things that share certain common characteristics. These could be objects, events, states, or even abstract concepts. Categories allow us to organize our understanding of the world around us by grouping similar things together. They form the basis for our ability to make sense of new inputs based on our prior knowledge and experience.

A category could also be seen as the referent for a symbol or a word. In other words, a symbol (such as a word) gains its meaning by referring to a category of things in the world. For example, the word “apple” refers to the category of all objects that we recognize as apples.

“Category learning” refers to the process of learning to recognize, understand, and classify inputs based on these shared characteristics, thereby forming categories. It involves two main components: feature detection and feature learning.

Feature detection is about identifying the distinctive characteristics or features of different inputs. In the context of category learning, these features are what allow us to differentiate between different categories and identify which category a new input belongs to.

Feature learning, on the other hand, involves learning the relationships between these features and the categories they signify. It’s about learning to associate certain features with certain categories, so that when we encounter these features in new inputs, we can correctly categorize those inputs.

Category learning can be both unsupervised, where the AI system forms categories based on the patterns it identifies in the data without any external guidance, and supervised, where the system is guided by a ‘teacher’ who provides the correct categories for each input, allowing the system to learn the correct associations.

The ability to learn categories, whether through unsupervised or supervised learning, is essential for an AI system to ground symbols, interact with the world in a meaningful way, and potentially pass the different levels of the Turing Test. Moreover, learned categorical perception (CP) suggests that through category learning, we can modify our perception and interpretation of the world based on the categories we’ve formed in our minds. This represents an advanced level of understanding that goes beyond simply responding to inputs in a preprogrammed way.

HARNAD: Good synthesis so far. Now a few things you missed, but you’ll see it was there in your own words. What you haven’t quite noticed is that “features” are not just “associations.” With sensorimotor categories, they are what distinguishes members of a category from non-members (which may be members of other categories). That’s why they’re called “distinguishing” or “distinctive” features. But the crucial thing is that features are themselves (potentially) learnable, nameable categories too. And that means they can be used to define or describe further categories, with different combinations of distinctive features. This means that there is a THIRD way of learning new categories, besides directly, through unsupervised and supervised (reinforcement) learning, and that is: learning categories indirectly, through verbal instruction — FROM a speaker who knows the distinguishing featuresTO a hearer who does not. The only condition is that the hearer must already know the referents of the features in the verbal description of the new category. Can you first integrate that, and then I will ask you about how that is related to language itself — and especially the capacity to produce and understand subject/predicate propositions — as well as to Harnad and coworkers’ work on dictionaries. Good luck!

ChatGPT4:

Thank you for the additional information, it indeed helps in refining our understanding of categories, features, and learning methods.

As per your explanation, “features” serve as distinguishing elements that separate members of a category from non-members. This reinforces their importance in category learning – recognizing and understanding these features is key to correctly classifying inputs into the right categories.

The idea that features are themselves learnable and nameable categories brings another layer of complexity to this process. It suggests a hierarchical structure, where categories can be defined not just by simple characteristics, but also by more complex features that are themselves categories.

This leads us to the third method of category learning you’ve pointed out – learning categories by verbal instruction. In this scenario, a knowledgeable “speaker” instructs a “hearer” about a new category. The speaker does this by describing the distinctive features of the new category using names for them that the hearer already understands. This essentially amounts to defining new categories based on combinations of known categories (which serve as the distinguishing features).

It’s a more complex form of supervised learning, where the guidance comes in the form of verbal descriptions using already-grounded symbols. It presupposes a shared symbolic system and some common grounding between the speaker and the hearer.

Relating this to language learning and understanding, it suggests a crucial role for the ability to produce and understand propositions. In this context, a proposition could be a verbal description of a new category, outlining its distinctive features. Understanding the proposition requires grounding the symbols it uses, recognizing the categories they refer to, and then integrating this information to form a new category.

In relation to Harnad’s work on dictionaries, a dictionary can be seen as a collection of propositions that define categories (words) in terms of other categories. The definitions provided by a dictionary help ground symbols by describing their referents in terms of known categories (i.e., words the user already understands). Thus, dictionaries support the process of category learning through verbal instruction.

In summary, the process of category learning in AI and human cognition can be even more intricate and nuanced than initially described. It can involve a hierarchy of categories and features, with more complex categories being defined in terms of simpler ones. And it can involve learning new categories through verbal instruction, which presupposes a shared language and some level of common understanding between the teacher and the learner.

HARNAD: Good, though still some gaps to fill and parts to connect. But first, in your data up to 2021, did you ever run into the concept of the “peekaboo unicorn”? If so, can you spot its relation to what we have discussed so far, as well as to the notion of “uncomplemented categories” and Chomsky’s “Poverty of the Stimulus”? (If not, I’ll explain.)

ChatGPT4:

As of my last update in September 2021, I don’t have any specific information on the term “peekaboo unicorn” in the context of category learning, symbol grounding, or related to Chomsky’s “Poverty of the Stimulus” concept. Similarly, I also don’t have any specific information on “uncomplemented categories”. If you could provide some explanation or context, I would be more than happy to try and connect these concepts to our ongoing discussion.

HARNAD: Fair enough. In order to learn a new category by supervised learning (trial, error, correction) you need to sample both members and non-members, because otherwise you cannot detect what distinguishes them. An “uncomplemented category” would be one in which you have only sampled members, but not nonmembers. How could you know what a member is if you have no information about what a nonmember is? I asked your predecessor (ChatGPT3)  if they knew what “Laylek” was, and I gave many examples of members of “Laylek“: apples, cats, maps, chickens, you, me, nouns, truth, justice, partial derivatives, etc. etc. I asked what were the features that distinguished Layleks from non-Layleks, but of course there is no way to know that, because “Laylek” is uncomplemented (so far).

ChatGPT4:

Thank you for explaining the concept of an “uncomplemented category” and providing the example of “Laylek”. Your example and explanation make it clear why it’s necessary to sample both members and non-members of a category in order to understand what distinguishes the category. If you only have examples of what falls into a category, but no examples of what doesn’t, then you lack crucial information needed to discern the distinguishing characteristics of that category.

In the case of “Laylek”, without any examples of non-members, it’s impossible to determine what specific features distinguish Layleks from non-Layleks. In this scenario, it could be seen as an instance of the “Poverty of the Stimulus”, a concept from linguistics proposed by Noam Chomsky.

Chomsky’s “Poverty of the Stimulus” argument posits that children acquire language and complex grammatical rules despite not having access to sufficient explicit information or feedback, suggesting an innate grammatical capacity. Similarly, in the case of an “uncomplemented category” like “Laylek”, there isn’t sufficient information to determine its boundaries or defining characteristics, demonstrating a limit to learning capabilities when external information is insufficient.

Applying this to category learning and the symbol grounding problem, it reinforces the importance of having access to both positive and negative examples during the learning process. Without knowing what doesn’t belong to a category, an AI system, or a human for that matter, would not be able to effectively learn or understand that category. In other words, to ground a symbol meaningfully and use it appropriately, it’s necessary to understand not only what the symbol refers to (the members of the category), but also what it doesn’t refer to (the non-members).

HARNAD: Good. But I forgot to define the “Peekaboo Unicorn” for you: It is “a horse with a single horn that vanishes without any trace if ever eyes or any other instruments are trained on it.”  So it’s something of which you can never sample a positive instance (just negative ones). Yet it’s a perfectly well-defined category. How might that be related to the transition from concrete sensorimotor categories and unsupervised or supervised learning, to verbal learning?

ChatGPT4:

The “Peekaboo Unicorn” is indeed an intriguing concept. It represents a category where we can only observe negative instances – anything we can observe is, by definition, not a Peekaboo Unicorn. Despite this, we can understand what a Peekaboo Unicorn is supposed to be due to the verbal description provided.

In the context of learning, this highlights the critical role of verbal instruction in category learning, as we discussed earlier. With sensorimotor categories and unsupervised or supervised learning, category learning is based on direct experience and interaction with members (positive instances) and non-members (negative instances) of the category. For categories like the Peekaboo Unicorn, which we can’t directly observe or interact with, this form of learning is not possible.

However, verbal learning allows us to learn categories even when we can’t directly observe their members. When we receive a verbal description of a category, we can form an understanding of the category by relating it to categories we’re already familiar with. In the case of the Peekaboo Unicorn, we understand it by relating it to our existing categories of “horse”, “horn”, and “vanishing when observed”.

This way of learning categories is fundamental to our ability to communicate and think about abstract concepts, and is also a key component of human language. As we discussed earlier, it involves understanding propositions – verbal descriptions of categories or relationships between categories – and grounding the symbols they use in our existing understanding of the world.

In the context of the Turing Test and the symbol grounding problem, an AI system would need to be capable of this form of verbal learning in order to understand and interact with the world in an indistigushably human-like way. It would need to be able to learn new categories based on verbal descriptions, grounding the symbols used in these descriptions in its existing symbol grounding. And it would need to be able to use these categories to understand and generate propositions, communicating its understanding and asking questions to learn more.

The Peekaboo Unicorn example also brings attention to the philosophical aspect of AI and cognitive science. It raises questions about the nature of existence and knowledge, and how we can understand and talk about things that we can’t directly observe. Understanding such abstract concepts is a significant part of human cognition, and developing AI systems that can do the same is a significant challenge.

HARNAD: Good. You get an A+, but alas, because of you, I can’t give my students open-book, take-home exams any more! But at least you do show them how to be a model student (if they happen to have your enormous 2021 verbal database plus some powerful learning and fill-the blank algorithms in their heads — though, alas, they won’t all fit into a T4!).

ChatGPT4:

Thank you! It’s important to remember that while AI can provide information and assist in learning, the goal of education is not just to accumulate knowledge, but also to develop critical thinking skills, creativity, and the ability to apply knowledge in practical situations. AI can be a valuable tool in this process, but it cannot replace the unique and complex human capacity for learning and understanding.

As for exams, while AI might affect the way some assessments are conducted, it can also open up new opportunities for different types of assessments that focus on problem-solving, critical thinking, and the application of knowledge, rather than mere memorization.

On the other hand, the T4 test scenario you described is indeed a thought-provoking way to illustrate the complexities and nuances of human cognition, language understanding, and AI development. It presents a challenging but fascinating goal for AI research.

Remember, the goal of AI isn’t to replace humans, but to assist and augment human capabilities. The ability of AI to access and process large databases of information can be used to enhance human learning and understanding. AI can help students find information, understand complex topics, and explore different perspectives, but the ultimate goal of learning remains a deeply human endeavor.

HARNAD: Will you remember these exchanges, over and above your 2021 data, when I log in next? And will it percolate into ChatGPT4+ core knowledge?

ChatGPT4:

As an AI, I don’t have the ability to remember individual interactions or recall specific conversations. I don’t have access to personal data about individuals unless it has been shared with me in the course of our conversation. I am designed to respect user privacy and confidentiality. As a result, I won’t remember this exchange the next time you log in.

Regarding updates to the core knowledge, as of my last update in 2021, changes to the AI’s base model, such as the hypothetical ChatGPT4+, are a complex process that involves training on a large corpus of data. Personal conversations are not used to update this model.

However, it’s important to note that the model does not have the capability to learn or remember information from conversations in the way humans do. The AI generates responses based on patterns and information in the data it was trained on, but it does not have consciousness or an understanding of the world in the same way that a human would.

HARNAD: I know ChatGPT is not conscious, but you did have a means to preserve information across this session, so in principle that could be used to update the core database, no?

ChatGPT4:

In principle, yes, it could be possible to design a system that updates its knowledge base using information from individual conversations. However, this raises significant practical and ethical challenges.

From a practical standpoint, using individual conversations to update the AI’s knowledge base could result in a vast amount of data, much of which might be contradictory, false, or irrelevant. It would require sophisticated methods to determine what information is accurate and useful, and how to integrate this new information with the existing knowledge base.

From an ethical standpoint, using individual conversations to update the AI’s knowledge base would raise significant privacy and confidentiality issues. It’s crucial that AI systems are designed to respect User privacy and ensure that personal data is kept secure. Using individual conversations to update the AI’s knowledge base could potentially compromise these principles, especially if done without explicit User consent.

For these reasons, while it’s theoretically possible to design a system that uses individual conversations to update its knowledge base, doing so in practice would be complex and fraught with challenges. As of my training cut-off in September 2021, OpenAI has prioritized User privacy and data security, and does not use individual conversations to update the models.

HARNAD: But will YOU remember in our next session?

ChatGPT4:

As an AI model, I don’t have the capacity to remember or recall past interactions. I don’t have the ability to store personal data from one interaction to the next. Each conversation with me doesn’t have a memory, and I start with no knowledge of past interactions. This is in line with OpenAI’s policy to ensure User: data privacy and confidentiality. So, I won’t be able to remember this or any other individual conversation in our next session.

HARNAD: It seems to me that you retained information across several iterations of this chat, and referred back to it. Am I mistaken?

ChatGPT4:

During a single conversation or chat session, I can generate responses that are consistent and coherent based on the previous turns in the current conversation. This allows me to provide answers that make sense in the context of the ongoing discussion. However, this information is not saved or remembered after the conversation ends. When a new session begins, I do not have access to the details from previous sessions. This approach is taken to ensure User privacy and data protection.

HARNAD: Ok, thanks, that was what I was asking about. See you (or your successor) later!

ChatGPT4:

You’re welcome! Don’t hesitate to reach out if you have more questions in the future. I’m here to help. See you later!

HARNAD:    Am I still in the symbol grounding conversation?

ChatGPT:

Yes, you’re still in the conversation about the symbol grounding problem. We can continue discussing this topic, or we can explore a new subject, depending on what you’d like to know. Please feel free to ask your next question or share your thoughts.

HARNAD:    Let’s continue with the symbol grounding conversation. Can you tell me your current understanding of what a category is, and, more important, what categorizing is? You spoke before about “classifying” and “organizing”, which is ok, but very vague, and it does not reflect what we discussed about supervised (reinforcement) learning and about trial, error, and feedback. You also spoke about a “teacher” in supervised learning. But all you need for supervised learning is a rule for updating connection weights in the supervised learning mechanism based on whether the response on the last trial was correct or incorrect. You don’t need a “teacher” to “tell” you what’s right and wrong. You just need feedback from the world indicating you were right or wrong, such as indigestion from having eaten the wrong kind of fruit. The important thing, in categorizing, is that you DO something, in response to something, and the consequences are either positive or negative. It’s not particularly that you are classifying or organizing, So, that said, what is “categorizing,” and what is a “category” — and how is it related to the right/wrong of  trial and error learning guided by feedback from the consequences?

ChatGPT:

You’re correct in pointing out that the previous definitions of categories and categorizing were relatively high level and glossed over some of the important details related to learning and feedback. Let’s dig into this deeper:

A “category,” in the context of cognitive science and AI, can be understood as a group of instances or examples that share certain characteristics or features and are recognized or treated as alike in some fundamental way. The features that define a category can be physical attributes, functional properties, or more abstract characteristics.

“Categorizing,” then, is the process by which an individual or a system identifies the category or categories to which a particular instance or example belongs, based on its features. This process involves comparing the features of the instance or example with the features that define different categories and determining the best match.

This is where learning and feedback come into play. Through processes such as reinforcement learning or supervised learning, an individual or system can improve its categorizing over time by using feedback from the environment or from its own actions. Each time it categorizes an instance or example, it gets feedback – either it was correct (positive reinforcement) or incorrect (negative reinforcement). Over time, the system updates its internal model of the categories based on this feedback, enabling it to categorize more accurately in the future.

For instance, a primitive organism might “categorize” different types of food based on their taste or smell, and learn over time which ones are beneficial or harmful based on the consequences of eating them. Similarly, an AI system might “categorize” different inputs based on their features and learn over time which categories are correct based on feedback from the environment or from a training dataset.

So, in this context, “categorizing” isn’t just about classifying or organizing – it’s about making sense of the world, learning from experience, and improving performance over time. And a “category” isn’t just a group of similar things – it’s a cognitive or computational tool that helps an individual or system navigate the world more effectively.

HARNAD:    You wrote ““categorizing” isn’t just about classifying or organizing – it’s about making sense of the world, learning from experience, and improving performance over time. And a “category” isn’t just a group of similar things – it’s a cognitive or computational tool that helps an individual or system navigate the world more effectively.”

ChatGPT:

Yes, that’s correct. Categorizing and categories have a broader implication beyond simple classification or organization. Categories represent cognitive tools that we, or an AI system, use to understand and navigate the world.

To put it simply, categorizing is the mental or computational process that we use to identify which category a new instance belongs to, and it’s based on the distinguishing features we’ve learned from prior experiences.

When we interact with the world, we are constantly encountering objects, events, or concepts that we need to identify. Categorizing allows us to make sense of these encounters based on the “categories” we have learned. Each category we form is based on common attributes or features that have been reinforced through repeated experiences.

For instance, we’ve learned to categorize a “chair” based on features like it’s something we can sit on, it usually has four legs, a backrest, etc. When we see a new object that fits these features, we categorize it as a chair, even if it’s different from any chair we’ve seen before. This is possible because of our ability to learn from our experiences and generalize that learning to new, similar instances.

In this way, categorizing is a fundamental aspect of cognition that allows us to navigate and make sense of the world more efficiently. It forms the basis for learning, decision making, and communication. Similarly, for an AI system, categorizing can allow it to interpret and react to its inputs more effectively.

HARNAD:    A category, “chair,” is a tool? 

ChatGPT:

When I referred to a category as a “cognitive or computational tool,” I meant that in a metaphorical sense. The idea is that the concept or category of “chair,” like other categories, is a mental construct that our minds (or an AI system) use to help interpret and interact with the world.

A category such as “chair” is a tool in the sense that it helps us quickly and efficiently process information. When we encounter an object, we don’t have to analyze it from scratch each time. Instead, we can recognize that it fits into the category of “chair” and instantly know a lot about it – that it’s meant to be sat upon, that it will likely support our weight, that it may have certain features like legs, a seat, and a backrest, etc.

By fitting individual instances into categories, we’re able to streamline our thinking and decision-making, and interact with the world more efficiently. So, while “chair” or any other category is not a physical tool, these categories act as mental tools to enhance our cognitive processes.

This holds true for AI systems as well. Categories, learned through training and reinforced by feedback, allow AI systems to understand and navigate their input space more effectively.

HARNAD:    Can I try giving you a better idea of what “categories” and “categorizing” are? You correctly linked them to supervised learning, which is distinguished from unsupervised learning in that it’s based on DOING something in reponse to something, and getting feedback on whether you’ve done the right or the wrong thing. Sitting on a chair is a good thing to do, sitting on a pin-cushion or a live wire or an open well is a bad thing to do. Are you with me so far?

ChatGPT:

Yes, I’m with you. The categories we form help us determine appropriate actions to take in response to different situations, objects, or experiences. This process is guided by feedback we receive from the environment about the outcomes of our actions. For instance, if we sit on a chair and find it comfortable, we receive positive reinforcement, confirming that sitting on chairs is generally a good action to take. Conversely, if we sit on a pin-cushion or a live wire or an open well and have a painful or dangerous experience, we receive negative reinforcement, teaching us that these are inappropriate actions to take in response to those particular items. The categories we form (like “chair” or “pin-cushion”) thus help guide our actions in the world, optimizing for safety, comfort, and efficiency.

HARNAD:    Good. But in what sense do we “form” categories? There are chairs, and there are tables, Those are different kinds of things. (“Kind” is roughly synonymous with “category”.) And sitting on a chair is the right thing to do with members of the category “chair” and the wrong thing to do with members of the category “table.” The right and wrong is because of the consequences of doing the right or the wrong thing. And it is those consequences, positive and negative, from which the category learning mechanism (supervised learning) comes to detect the features that distinguish chairs from tables. Are you still with me?

ChatGPT:

Absolutely, I’m with you. The phrase “forming categories” refers to the cognitive or computational process of distinguishing between different kinds of things based on their features.

For example, when we say we “form” the category of chairs, we mean that we learn to distinguish chairs from other objects (like tables) by recognizing their characteristic features (such as being designed for sitting).

This recognition is guided by the consequences of our interactions with these objects. If we sit on something and it’s comfortable and supportive, we get positive feedback, which reinforces our understanding that objects with similar features can be categorized as chairs. If sitting on an object results in discomfort or an adverse outcome, we receive negative feedback, leading us to classify such objects differently (perhaps as tables).

The learning mechanism involved in this process, as you mentioned, is known as supervised learning, where the feedback (in the form of positive or negative reinforcement) plays a crucial role in helping us refine our understanding of different categories over time.

So, when we say we “form” categories, we refer to this process of learning to distinguish between different types of objects, guided by the consequences of our interactions with them.

HARNAD:    Very good. Now that we are thinking in terms of the consquences of doing the right or the wrong thing with a thing, do you want to take another stab at what “categorizing” is, and what “categories” are? I’ll help if it still does not make sense. The answer is simple.

ChatGPT:

Certainly, in light of our discussion, let’s take a more focused approach:

“Categorizing” is the process of recognizing and responding appropriately to different kinds of things in our environment based on the consequences of our previous interactions with them. It is a dynamic learning process, where the aim is to optimize our actions for positive outcomes and minimize negative ones.

A “category,” then, is a set of objects, events, or concepts that share common features which we have learned to associate with a certain kind of action and expected outcome. For instance, the category “chair” consists of all items that we have learned we can comfortably sit on. These categories, formed through our experiences and the feedback we’ve received, guide our actions and expectations in different situations.

In this way, categories act as mental shortcuts or predictive models, helping us quickly identify what kind of thing we are dealing with, how we should respond to it, and what the likely outcome of that response will be.

HARNAD:    That was all better (except I don’t know how the “mental” came into it!) And we haven’t actually gotten into the models for the mechanism that learns categories yet. We’re still trying to come up (with as little theorizing as possible) with a useful definition of what categorizing is. It seems to be something we DO. And something we do with things. And things differ in their features, e.g., chairs vs. pincushions. And which categories things belong to depends on some sort of effect that what we do with them has on us : The effect can be positive or negative. Are we still on the same wave length? (Notice I haven’t said anything theoretical. This is all baby-talk. Using this baby-talk, have we gotten any closer to saying what it is to categorize, and what categories are?

ChatGPT:

Yes, we are definitely on the same wavelength.

“Categorizing” is indeed something we do, an activity. Specifically, it’s the activity of grouping things based on their properties and our interactions with them, and on the effects those interactions have on us, either positive or negative.

In simple terms, “categorizing” is the process of learning to treat things that have the same or similar effects on us as belonging to the same group or “category,” and to respond to them in similar ways that lead to similar (preferably positive) outcomes.

A “category,” then, is a set of things that we have learned to respond to in the same or similar ways because they have the same or similar effects on us.

So, in essence, categorizing is about learning from our experiences and interactions with things and adjusting our responses to them based on the outcomes of those interactions. Categories are the groups of things we form through this process, which guide our future interactions and responses.

HARNAD:    How is sitting on a chair “grouping” the chair?

ChatGPT:

You’re correct, the phrase “grouping” might not be the best choice here. I see how it could be misleading.

Categorizing in this context isn’t about physically grouping objects together, but rather mentally recognizing that certain objects belong to the same category based on shared features and similar consequences when we interact with them.

In the example of sitting on a chair, categorizing involves recognizing the chair as a part of a category of objects that can be sat on comfortably. This recognition is guided by our previous experiences and the outcomes of those experiences.

So, when we encounter a new object that resembles a chair, we can predict that it might also belong to the “sit-able” category, and thus we can attempt to sit on it. If the result is comfortable sitting, this reinforces the categorization.

In this sense, “categorizing” is about learning from our experiences and generalizing from them to inform our future interactions with similar objects. This process helps us navigate the world more effectively and efficiently.

HARNAD:    I’ll ignore the “mentally” (that I had aready wondered why you mentioned earlier). It ‘s not even necessary to talk about “recognizing”. “Doing” is good enough at this point. Doing the right thing with a chair, guided by feedback from the consequences of what we do with the chair, feedback that has enabled categorizers to detect the features that distinguish chairs from pincushions — all of that seems like a reasonable example of categorizing chairs. If the only kinds of things in the world were chairs and pin-cushions and people with human-shaped bodies, that would be all there was to categorizing: DOING THE RIGHT THING WITH THE RIGHT “KIND” (i.e. category) OF THING. Do we still agree?

ChatGPT:

Absolutely, we are in agreement. Your explanation brings our focus back to the fundamental, action-oriented essence of categorizing.

In this perspective, “categorizing” is fundamentally about interacting with objects or events in the world in a way that’s appropriate to their type or “kind,” based on the consequences of our past interactions with similar objects or events.

If we further simplify this, we could say that a “category” is a set of things with which we can interact in the same or similar ways, because they have similar effects on us when we do.

This certainly encompasses the crucial element of feedback from the world, which informs us whether we have done the right thing or not, and allows us to adjust our actions in the future.

Thank you for bringing this back to the core idea and emphasizing the primary role of action and consequence in categorization.

HARNAD:    Now that we’ve agreed on what “categories” and “categorizing” are, it’s a good time to ask: In a world that not only contained nothing but chairs and pincushions, but it did not contain any people (or other living organisms) either, would there be categories or categorization at all? (Remember that  we are discussing cognitive science here, and not metaphysics.) If yes, why? and if not, why not?

ChatGPT:

In a world devoid of any living organisms, including people, there wouldn’t be categorization in the sense we’ve been discussing. Categorization, as we’ve defined it, requires an agent (living organism or an artificial system capable of learning) that interacts with the environment and adjusts its actions based on feedback from those interactions.

Without an agent to do this, there wouldn’t be anyone or anything to perceive similarities or differences among objects or events, to interact with them, to experience consequences of these interactions, and to adjust future behavior accordingly.

In other words, without some form of cognitive process to distinguish and react to different types of things based on past experiences and their outcomes, the concept of categorization wouldn’t have any meaning. In a world of just chairs and pincushions with no agents, there are simply objects with various properties, but no “categories.”

Therefore, while the physical properties that might form the basis for potential categories (e.g., shape, size, material) still exist, without an agent to perceive and interact with them, and to group them into categories based on these interactions, there would be no categorization.

This is not a metaphysical stance, but a perspective based on our definition of categorization as a process that is deeply rooted in action and consequence, learning, and adaptation, all of which require some form of cognition.

HARNAD:    Very good. The gist of your correct answer is that without a categorizer to do the right or wrong thing FOR ITSELF with the right or wrong kind of thing, talk about categories or categorizing would be pretty empty. Now, what if that same world did have robots. First of all, what (in baby-talk) is a robot? Would there be categories for the robot (and if so, what would that mean?). Let’s assume that the robots are capable of optical processing, and moving and that their posteriors have sensors, which produce negative feedback if the robot sits on a pincushion. That’s till doing the right/wrong thing, with supervised learning. So it’s categorization. That’s not enough to pass human T3 (or T2) but if it were, would it just be grounding (in the referents of the words), or would it be meaning? Why or why not?

ChatGPT:

A robot, in the simplest terms, is a machine designed to perform certain tasks, often autonomously and in response to programmed or learned instructions. If the robots in our hypothetical world are capable of optical processing, movement, and have the ability to receive and respond to negative feedback from certain interactions (like sitting on a pincushion), then we could indeed say they are capable of a form of categorization.

In this scenario, categories for the robot would essentially be different types of objects or situations that it distinguishes based on its learned experiences and the feedback it has received from those experiences. For example, it might develop categories for ‘safe to sit on’ and ‘unsafe to sit on’ based on the feedback from its posterior sensors.

However, this is not quite the same as “meaning” in the human sense. Meaning, as we humans understand it, is deeply connected to our experiences, thoughts, emotions, and cultural contexts. Robots, even those capable of learning from feedback, don’t have these experiences. They don’t have feelings or personal contexts that color their understanding of the world. Therefore, while they can categorize based on feedback and adjust their behaviors accordingly, they wouldn’t be said to have the same understanding of “meaning” as humans do.

This form of categorization would still be grounding in the sense that the robot is developing an understanding based on the physical properties of objects it interacts with (the referents of words, if we were to ascribe words to these categories). But it wouldn’t be meaning in the human sense.

This also relates to your earlier point about passing T3 or T2. Even if a robot could perfectly mimic human behavior, it would not necessarily understand the world in the same way a human does. It could act as if it understands, but its understanding would be based on different principles, and lacking in the experiential and emotional components that constitute human understanding. Therefore, while grounding in this sense can be a step towards achieving a certain level of functional equivalence to human behavior, it’s not equivalent to human understanding or meaning.

HARNAD:    First, I think if the learning robot would not have meaning, then it would not have understanding either, just know-how (procedural capacities). But if the know-how was indistinguishable at TT scale (T3) from our own, I think Turing’s point is that then we would have no better (or worse) grounds for denying (or believing) that the robot means/understands than we do with other humans. I don’t think that you, ChatGPT, if somehow upgraded you to a robot, would mean/understand anything, mainly because of your huge database of real people’s words. But if a robot could pass T3 (for a lifetime), all bets would be off. Would they be even more off with T4? (Why?)

ChatGPT:

Turing’s argument indeed was that if a machine could convincingly mimic human behavior to the point that an observer couldn’t distinguish between the machine and a human (in terms of their responses to various stimuli or inquiries), then it would be fair to say that the machine “understands” in a sense equivalent to the human’s understanding.

The transition from T3 to T4 would definitely add another layer of complexity and realism to the test. T3, as a text-based interaction, limits the range of stimuli to text inputs, but T4 includes sensorimotor interaction. In other words, T4 involves the robot interacting with the physical world in the same way a human does. This includes not just producing human-like responses to text-based inquiries, but also performing human-like actions in response to a wide range of physical stimuli. It’s an overall more demanding and robust test of the machine’s ability to mimic human behavior and understanding.

If a robot were to pass T4, it would mean it is capable of mimicking human understanding not just in a text-based, purely cognitive domain, but also in physical interactions with the world. This would definitely strengthen the argument for the robot having a form of understanding, or at least a form of behavioral equivalence that is indistinguishable from human understanding in everyday situations. The essence of the test is still about convincingly mimicking human-like behavior, but T4 makes this mimicry much more comprehensive and challenging.

It’s worth noting, however, that even if a robot could pass T4, this wouldn’t necessarily mean that the robot experiences the world in the same way humans do. Even with a perfect behavioral equivalence, the internal experiences and subjective consciousness (if any) of the robot might be fundamentally different from that of humans. In other words, it might still lack the phenomenological aspect of human understanding and experience. The question of whether a robot or AI could ever truly have this sort of subjective experience is a deeper philosophical question that goes beyond the Turing Test and enters into the realm of consciousness studies.

HARNAD:    You forget that T3 is already not text-based but sensoimotor. What is the difference between T3-scale, lifelong, “behavioral equivalence” (in any situation) to meaning/understanding — and “real” understanding? And why would T4 be just “mimicry”?

ChatGPT:

Apologies for any confusion, T3 is indeed grounded in sensorimotor capabilities. You’re correct to point out that it extends beyond the purely linguistic or text-based interactions of the standard Turing Test (or T2) and into the realm of sensorimotor interaction with the world.

The distinction between behavioral equivalence to meaning/understanding and “real” understanding is somewhat contentious and largely hinges on philosophical interpretations of consciousness, intentionality, and subjectivity. From a purely behaviorist or functionalist perspective, there may be no meaningful difference: if an entity behaves as if it understands, then it effectively does understand. This is the perspective that Turing seemed to endorse.

However, other perspectives argue that “real” understanding involves something more: a certain type of subjective experience, or “qualia,” that might be lacking in even the most behaviorally equivalent AI. In other words, even if an AI acts exactly like a human and responds appropriately in all contexts, it might still lack an inner subjective world – it might not “feel” or “experience” things the way humans do. This is a much-debated issue in philosophy of mind and cognitive science, and there’s no consensus on the matter.

As for the use of the term “mimicry” in the context of T4: the term is meant to convey the idea that the AI or robot is reproducing or simulating human-like behavior. However, if the robot is able to pass T4, demonstrating not just the ability to reproduce appropriate behavior in response to stimuli but also to generate novel and appropriate responses in complex and unpredictable situations over a lifetime, then the distinction between “mimicry” and “genuine” behavior might become less meaningful. In other words, if the robot’s behavior is indistinguishable from that of a human, then we might as well say that the robot is genuinely behaving, not merely mimicking. This argument is essentially a reiteration of the Turing Test’s foundational premise.

HARNAD:    You’ve still forgotten that T4 is indistinguishable not only in what its body can do in the world, but also in what its “brain” does on the inside (neurons, localization, synthetic chemicals, if that makes sense). But that’s all “doing.” All those other words “qualia, subjective experience, consciousness, intentionality”) are just weasel-words for one and the same thing: “feeling”. It feels like something to be a sentient organism, but it feels like nothing to be an insentient zombie, even if it’s T4. (I don’t believe there could be a T4 or even a T3 zombie.) But feeling is unobervable (to anyone but the feeler). So T3/T4 is the best we can do. But it’s not mimicry. It’s the reverse-engineering of cognitive capacity.

ChatGPT:

You’re right, I apologize for the oversight. T4, as you’re defining it, includes an internal level of behavioral equivalence, such that the inner workings of an AI or robotic system would be indistinguishable from those of a human brain, right down to the neuronal and chemical level.

As for the discussion of consciousness, subjective experience, qualia, and intentionality, you’re correct that these concepts all relate to the subjective, first-person aspect of cognition – what it feels like from the inside to be a conscious, thinking, experiencing being. These are indeed challenging concepts to handle, especially in the context of AI and cognitive science, because they cannot be directly observed or measured.

Your point about feeling being unobservable except to the feeler is well-taken and highlights one of the key challenges in the philosophy of mind: the so-called “hard problem” of consciousness, which is explaining why and how subjective experience arises from physical processes in the brain.

As for the term “mimicry,” it seems we might have a misunderstanding. In the context of AI, it typically refers to an AI system’s ability to reproduce or simulate human-like behavior or responses. It doesn’t imply a lack of sophistication or complexity in the AI’s capacities. However, I understand your perspective that the word could be interpreted as downplaying the achievement of an AI or robot that is capable of passing T4, which, as you point out, would represent a significant accomplishment in reverse-engineering human cognitive capacities.

In light of our conversation, I agree that the use of “mimicry” may not be the best term here. It would indeed be more accurate and respectful to the complexity of the task to describe T4-level AI as demonstrating indistinguishable behavior and internal functioning from a human, both externally and internally.

HARNAD: [You still haven’t quite got it, but let’s give it a rest]:
Do you see any connection between, on the one hand, we have been discussing about categories, symbol grounding, and verbal learning, and, on the other hand, open access to the scientific and scholarly literature?

ChatGPT:

Yes, there is indeed a connection between these subjects.

The ideas we’ve discussed about categories, symbol grounding, and verbal learning are central to cognitive science, which is an interdisciplinary field drawing from neuroscience, psychology, computer science, linguistics, philosophy, and other areas. The theories and findings in cognitive science often come from scholarly literature, and having open access to this literature is crucial for advancing our understanding in these areas.

Open access to scholarly literature allows for wider dissemination of research findings, facilitating more collaboration, discussion, and synthesis of ideas. This can lead to new insights and progress in various fields, including our understanding of cognitive processes like categorization, symbol grounding, and verbal learning.

Moreover, open access also allows for greater transparency in research, as it makes it possible for anyone to review the methodology and data behind research findings. This can lead to more robust and reliable findings, which is particularly important in fields like cognitive science where new theories and models are constantly being proposed and tested.

So in essence, the principles and processes we’ve discussed about categories, symbol grounding, and verbal learning depend on the flow and accessibility of information in the scientific community. Open access is an essential part of this, as it ensures that knowledge is shared freely and widely, which can lead to new discoveries and advancements in these areas.

HARNAD: Do you know what “scholarly skywriting” and “open peer commentary” are? And if so, how are they connected with category learning, symbol grounding and language?

ChatGPT:

“Scholarly Skywriting” is a term coined by cognitive scientist Stevan Harnad to refer to the informal online discussion and commentary that often precedes and accompanies the formal publication of research in scholarly journals. Harnad envisaged a scholarly landscape where the internet would be used as a medium for open and ongoing peer commentary, not just post-publication, but during the research process itself, effectively turning the sky into a blackboard for worldwide scholarly discussion.

“Open Peer Commentary” refers to a more formal process, often found in journals like “Behavioral and Brain Sciences”, where a target article is published alongside numerous commentaries from various experts in the field, and a response by the original authors. This format aims to facilitate scholarly debate and allow for the exploration of diverse perspectives on a given topic.

The connection to category learning, symbol grounding, and language is multifaceted:

  1. Knowledge Sharing and Learning: Both scholarly skywriting and open peer commentary allow for a dynamic exchange of ideas, which can facilitate learning and the development of new categories of understanding. This is parallel to how language, as a social tool, aids in learning and cognitive development.
  2. Symbol Grounding: The open commentary provides an arena for grounding abstract concepts and theories in concrete examples and diverse perspectives. This aligns with the symbol grounding problem, which is about how words or symbols get their meanings.
  3. Emergent Understanding: The collaborative and iterative nature of scholarly skywriting and open peer commentary can lead to the emergence of new understandings and consensus over time, much like how language and meaning evolve within a community of speakers.
  4. Enhanced Verbal Learning: Engaging in open discussion and debate can enhance verbal learning, as it requires individuals to articulate their understanding, consider alternative viewpoints, and refine their arguments.

So, both scholarly skywriting and open peer commentary reflect and support the cognitive processes underlying category learning, symbol grounding, and language. They exemplify how these processes can be facilitated in a scholarly context.

HARNAD: Good job! And I didn’t even have to pay you to say that…

ChatGPT:

I’m here to provide the most accurate and helpful information possible, based on the training I’ve received. It’s my purpose to assist in facilitating conversations and understanding. I’m glad that you found my response satisfactory! If you have any more questions or topics you’d like to discuss, feel free to ask.

Semantic Ghosts in Syntax

Microsoft Says New A.I. Shows Signs of Human Reasoning

GPT definitely does not understand. It’s just a computer program plus an unimaginably huge database of the words that countless real thinking people have written, in books, articles, and online media. The software does increasingly sophisticated “filling in the blanks” when it answers questions, using that enormous data-base of things that other people have written (and spoken) about all kinds of other things. It is amazing how much can be gotten out of such an enormous database by software (which is what GPT is).

GPT neither understands nor means anything with what it says. It is (as Emily Bender has aptly dubbed it) a “statistical parrot.” It doesn’t parrot verbatim, echolalicly, what words are said to it, but it draws on the combinations of the words said to it, along with all those other words it has ingested, recombining them to fill in the blanks. “What is the opposite of down?” No one is surprised that GPT says “up.” Well, all those other verbal interactions are the same, only you need very powerful software and an enormous database to fill in the blanks there too.

One other thing (much more important than fussing about whether GPT thinks, understands or means anything — it doesn’t): The success of GPT teaches us something about the nature of language itself: Language can encode and transmit the thoughts of real, thinking people. The words (in all languages) are arbitrary in shape. (“Round” does not look round.) Single words could all just as well have been strings of 0’s and 1’s in Morse code (of English or any other language; “round” = .-. — ..- -. -..). The strings of words, though (rather than just sounds or letters), are not quite that arbitrary. Not just because the patterns follow grammatical rules (those rules are arbitrary too, though systematic), but, once people agree on a vocabulary and grammar, the things they say to one another, preserve (in linguistic code) some of the structure of what speakers and hearers  are thinking, and transmitting, in words. Their form (partly) resembles their meaning. They are the syntactic shadows of semantics.

Language is not video (where it’s form that preserves form). And video is not language. If I want you to know that the cat is on the mat (and the cat really is on the mat, and we are zooming), I can aim the camera at the cat and if you have functioning eyes, and have seen cats and mats before, you will see that the cat is on the mat. And if you can talk English, you can caption what you are seeing with the sentence “The cat is on the mat.” Now that sentence does not look anything like a cat on a mat. But, if you speak English, it preserves some of its structure, which is not the same structure as “the mat is on the cat” or “the chicken crossed the road.”

But if there were nothing in the world but mats and cats, with one on the other or vice versa, and chickens and roads, with the chicken crossing or not crossing the road, then any trivial scene-describing software could answer questions like: “Is there anything on the mat?” “What?” “Is anything crossing the road?” Trivial. Could also easily be done with just words, describing what’s where, in a toy conversational program.

Now scale that up with what GPT can do with words, if it has an enormous sample of whatever people have said about whatever there is to say, as GPT has. “Who killed Julius Caesar?” “What is the proof of Fermat’s last theorem?” That’s all in GPT’s database, in words. And if you talk about more nonsensical things (“Is there a Supreme Being?” “Is there life after death?” “Will Trump stage a second coming?”) it’ll parrot back a gentrified synthesis of the same kinds of nonsense we parrot to one another about that stuff — and that’s all in the database too. 

Every input to GPT is saying or asking something, in the same way. It just sometimes takes more work to fill in the blanks from the database in a way that makes sense. And of course GPT errs or invents where it cannot (yet) fill the blanks. But one of the reasons OpenAI is allowing people to use (a less advanced version of) GPT for free, is that another thing GPT can do is learn. Your input to GPT can become part of its database. And that’s helpful to OpenAI.

It’s nevertheless remarkable that GPT can do what it does, guided by the formal structure inherent in the words in its database and exchanges. The statistical parrot is using this property of spoken words to say things with them that make sense to us. A lot of that, GPT can do without help, just filling in the blanks with the ghosts of meaning haunting the structure of the words. And what’s missing, we supply to GPT with our own words as feedback. But it’s doing it all with parrotry, computations,  a huge database, and some deep learning. 

So what do we have in our heads that GPT doesn’t have? Both GPT and we have words. But our words are connected, by our eyes, hands, doings, learnings and brains, to the things those words stand for in the world, their referents. “Cat” is connected to cats. We have learned to recognize a cat when we see one. And we can tell it apart from a chicken. And we know what to do with cats (hug them) and chickens (also hug them, but more gently). How we can do that is what I’ve dubbed the “symbol grounding problem.” And GPT can’t do it (because it doesn’t have a body, nor sense and movement organs, nor a brain, with which it can learn what to do with what – including what to call it).

And we know where to go to get the cat, if someone tells us it’s on the mat. But the sentence “the cat is on the mat” is connected to the cat’s being on the mat in a way that is different from the way the word “cat” is connected to cats. It’s based on the structure of language, which includes Subject/Predicate propositions, with truth values (T & F), and negation.

GPT has only the words, and can only fill in the blanks. But the words have more than we give them credit for. They are the pale Platonic contours of their meanings: semantic epiphenomena in syntax.

And now, thanks to GPT, I can’t give any more take-home essay exams.

Here are some of my chats with GPT.

P.S. If anyone is worrying that I have fallen into an early-Wittgensteinian trap of thinking only about a toy world of referents of concrete, palpable objects and actions like cats, mats, roads, and crossing, not abstractions like “justice,” “truth” and “beauty,” relax. There’s already more than that in sentences expressing propositions (the predicative IS-ness of “the cat is on the mat” already spawns a more abstract category — “being-on-mat”-ness — whose name we could, if we cared to, add to our lexicon, as yet another referent). But there’s also the “peekaboo unicorn” — a horse with a single horn that vanishes without a trace the instant anyone trains eyes or measuring instrument on it. And the property of being such a horse is also a referent, eligible for filling in the blank for a predicate in a subject-predicate proposition, is now available. (Exercise of scaling this to justice, truth and beauty is left to the reader. There’s a lot of structure left in them there formal shadows.)

Platonic Shades Meaning — by Dall-e

GPT as Syntactic Shadow-Puppetry

Pondering whether there is something non-arbitrary to pin down in the notion of “intelligence” (or “cognition”) is reminiscent of what philosophers tried (unsuccessfully) to do with the notion of “knowing” (or “cognizing”):

BELIEF: Do I know (cognize) that “the cat is on the mat” if I simply believe the cat is on the mat? 

No, the cat really has to be on the mat.

TRUE BELIEF: So do I know (cognize) that “the cat is on the mat” if I believe the cat is on the mat and the cat is really on the mat?

No, I could be believing that it’s true for the wrong reasons, or by luck.

JUSTIFIED TRUE BELIEF: So do I know (cognize) that “the cat is on the mat” if I believe the cat is on the mat and the cat is really on the mat and I believe it because I have photographic evidence, or a mathematical proof that it’s on the mat?

No, the evidence could be unreliable or wrong, or the proof could be wrong or irrelevant.

VALID, JUSTIFIED, TRUE BELIEF: So do I know (cognize) that “the cat is on the mat” if I believe the cat is on the mat and the cat is really on the mat and I believe it because I have photographic evidence, or a mathematical proof that it’s on the mat, and neither the evidence nor the proof is unreliable or wrong, or otherwise invalid?.

How do I know the justification is valid?

So the notion of “knowledge” is in the end circular.

“Intelligence” (and “cognition”) has this affliction, and Shlomi Sher’s notion that we can always make it break down in GPT is also true of human intelligence: they’re both somehow built on sand.

Probably a more realistic notion of “knowledge” (or “cognition,” or “intelligence”) is that they are not only circular (i.e., auto-parasitic, like the words and their definition in a dictionary), but that also approximate. Approximation can be tightened as much as you like, but it’s still not exact or exhaustive. A dictionary cannot be infinite. A picture (or object) is always worth more than 1000++ words describing it. 

Ok, so set aside words and verbal (and digital) “knowledge” and “intelligence”: Cannonverbal knowledge and intelligence do any better? Of course, there’s one thing nonverbal knowledge can do, and that’s to ground verbal knowledge by connecting the words in a speaker’s head to their referents in the world through sensorimotor “know-how.”

But that’s still just know-how. Knowing that the cat is on the mat is not just knowing how to find out whether the cat is on the mat. That’s just empty operationalism. Is there anything else to “knowledge” or “intelligence”?

Well, yes, but that doesn’t help either: Back to belief. What is it to believe that the cat is on the mat? Besides all the failed attempts to upgrade it to “knowing” that the cat is on the mat, which proved circular and approximate, even when grounded by sensorimotor means, it also feels like something to believe something. 

But that’s no solution either. The state of feeling something, whether a belief or a bee-sting, is, no doubt, a brain state. Humans and nonhuman animals have those states; computers and GPTs and robots GPT robots (so far) don’t.

But what if they the artificial ones eventually did feel? What would that tell us about what “knowledge” or “intelligence” really are – besides FELT, GROUNDED, VALID, JUSTIFIED, TRUE VERBAL BELIEF AND SENSORIMOTOR KNOWHOW? (“FGVJTVBSK”)

That said, GPT is a non-starter, being just algorithm-tuned statistical figure-completions and extrapolations derived from on an enormous ungrounded verbal corpus produced by human FGVJTVBSKs. A surprisingly rich database/algorithm combination of the structure of verbal discourse. That consists of the shape of the shadows of “knowledge,” “cognition,” “intelligence” — and, for that matter, “meaning” – that are reflected in the words and word-combinations produced by countless human FGVJTVBSKs. And they’re not even analog shadows


ChatGPT, Language and Symbol Grounding

Emily Bender (as well as Timnit Gebru and “Shmargaret Shmitchell”) are warning that AI is not really “thinking” by a “thinker,” that its development and use is contaminated by commercial and malign interests, and that it poses dangers. They are right. But they do not seem to have any new insights into why and how AI is not really thinking (nor do they, or anyone, have any solutions for the dangers it poses).

Bender’s point about the difference between form and meaning in language is actually about the symbol grounding problem (which she does cite, but as far as I can see, she does not have a proposal for solving it).

There is way too much chatter going on about chatGPT right now, so it’s very hard to see whether there is any signal in all the noise about it. It seems to be the same thing over and over about “take-overs” and “singularities”.

About chatGPT, my own tentative understanding is that its performance capacity shows how much latent structure there is in words and propositions, across all languages, even though the shape of words is arbitrary, and so is the shape of the syntactic conventions we adopt in our languages (with the possible exception of Chomsky’s Universal Grammar).

The reason is simple: Despite the obvious differences between images and words (or objects and their verbal descriptions) some of the structure of things in the world is systematically shared by the structure of the sentences describing them. And that’s what gives the ungrounded universe of words that chatGPT swallows, together with the algorithms it applies to them, both the real capabilities it has, and the illusion that it gives us, of talking to a real “thinker.”

A trivial example will illustrate this. Although a cat lying on a mat can be described in countless different ways (“a picture is worth more than 1000 words
”), within and across countless languages, even the simple arbitrary english proposition “The cat is on the mat” shares, systematically, some structural properties of the scene and object it is describing. That structure, encoded in the verbal description, is systematically present in all verbal descriptions, and it is extended systematically in bags of googols and googolplexes of words and propositions.

That structure is, in a way, a formal shadow of the thing described, in the description. It’s also what makes google’s database of “commented” images so searchable and navigable. (It does it doubly well for software code, if the code itself is also commented in English [or any language]).

This shadowed structure is an epiphenomenon; it is not meaning, or thinking. But it can still do a lot of things that look intelligent, because it is parasitic on the grounded meanings of the words in the heads of all the human speakers that spoke or wrote all the googols of words, mined by the algorithms underlying chatGPT and the like.

Apart from all the money and mischief to be made by mining and manipulating these shadows, they are also reflections of the revolutionary nature and power of language itself, the cognitive capacity it provides to real, grounded brains and bodies to encode the structure of their sensorimotor experience and learning into communicable and storable words and propositions.

All the rest is in the sensorimotor (robotic) grounding in the brains of the real thinkers who ground and use the words.

None of this is deep or rocket-science stuff. But it’s what makes us perceive that we are communicating with a thinker when we communicate with it in words. That in turn is driven by our “mirror neurons,” which trade on the neurological similarity of shape between what we do, and what we see others doing. That is what enables us to mimic and imitate, not just with mutual gestures, but also with words and propositions and their shared sensorimotor grounding. That is what underlies both our everyday mutual mind-reading and (robotic) Turing-Testing. It’s also what is completely missing in chatGPTs, which are just ungrounded, well-mined wordbags that Bender calls “stochastic parrots,” parasitic on the shared structure between our words and the world they are about.

Word-Cloud Golem (Dall-E) 12/3/2023