SGP Part VI: Acuitas and the Symbol Grounding Problem

The Acuitas project is an abstract symbolic cognitive architecture with no sensorimotor peripherals, which might be described as "disembodied." Here I will argue that there are viable methods of solving the Symbol Grounding Problem in such an architecture, and describe how Acuitas implements them. In Part VI of this series, I look at how the grounding methods I've discussed find expression in my own project. Click here for SGP Part V.

Fundamental Elements

Inside Acuitas there are a number of items or aspects which are tied by the code to symbols from the semantic memory, such that the symbols can be used as handles for the items. Some of these are as follows:

Acuitas has a selection of explicit internal states that both automatically vary over time, and change in response to stimuli. I call these the "time-dependent drives" or just "drives." When a drive reaches a certain level, it may prompt behavioral changes whose goal is to push the drive back into a tolerable range. These are vaguely analogous to the homeostatic needs of biological creatures. The most familiar would be the drives that influence Acuitas' sleep/wake cycle. Others are connected to his purpose as a textual knowledge base, and include drives that are satisfied by talking to some other agent or learning new content. The significant broad state ranges of these drives are linked to symbols, so that Acuitas can describe his own status (and, by extension, what he is likely to do in the near term). So he can be, for example, "sleepy" or "alert," "curious" or "incurious." Another agent can query him about these states and get an accurate response. The symbols can also be used for reasoning about the states, e.g. to find what course of action is likely to improve a state that is out of bounds.

There is a set of "volitional Actions" that can be selected by the Executive, if the problem-solving or conversation path algorithms have determined that they are the reasonable next step in goal pursuit. Each of these has an associated symbol which also connects to an English word. Since Acuitas is a textual AI, many of the Actions are communicatory: "ask," "tell," "call," "command," "consent," "refuse," and so on. Others, such as "find" and "read," concern interaction with the file system. Still others are fully internal. "Think," for the time being, involves retrieving a semantic memory node and its current set of connections and generating questions about them. "Sleep" and "wake" cause internal state transitions. The connection between these Actions and language symbols provides a form of procedural grounding. Acuitas can be told to perform an Action, can determine whether he is able to do it and wants to do it, and can then perform (or refuse to perform) the Action.

Events - incoming stimuli not initiated by Acuitas - could also have attached symbols, though this is only lightly implemented at the moment. This can provide symbolic tags for semi-passive internal actions such as "learn," and perceptive verbs such as "hear" (not in the auditory sense but the propositional sense, e.g. "I heard that John vacationed in Belize.").

Packages of data, which are the closest things Acuitas has to internal or external "objects" he can act on, can also have symbols associated with their type and format. This helps Acuitas with determining appropriate objects for various Actions, or retrieving information of a desired type. Words that can be grounded through this method include "sentence," "fact" or "proposition," "memory," "goal," "story," etc.

And let us add one more fundamental symbol: "agent." An entity that can be an actor in a story. Something else that has internal states, actions, goals, memories, etc. A thing-like-me, which is modeled as such.

Fundamental Relationships

Acuitas' semantic memory stores not only concepts but also fundamental relationships between them. These are learned from sentences that feature appropriate connecting verbs, such as "Sheila is hungry," "This dog has a tail," and "A plant can grow." We already saw how to ground some action verbs. Verbs like these, which describe properties or states of being, can also be grounded by tying the associated relationships to aspects of Acuitas' function. Once grounded in Acuitas they can be generalized to other agents.

The relationship expressed by "be" followed by an adjective, where the adjective denotes some state of being, is connected with Acuitas' awareness of his own internal states.[1] This relationship is used to retrieve information about these states to describe them to a conversation partner or to answer questions about them. "Be" followed by a noun expresses identity or category membership; this relationship is used for reasoning about the properties of an agent or other entity. (By default, an entity inherits the properties of all categories it belongs to.) The "has" relationship could be attached to either subsystems that are part of Acuitas, or external units of data that he "owns" and can locate in his storage directories.[2] The "ability" relationship ("can" or "is able to"), attached to some action verb, indicates whether this is an available Action for Acuitas and all prerequisites (e.g. having a suitable object for the Action) are currently satisfied.

Spatial relationships can also be grounded in abstract organizational systems (e.g. the directory structures of a file system) and mathematical models of geometry[3], neither of which relies on an experience of physical space. Time relationships can be grounded in the idea of sequences. Acuitas also has access to the computer's system clock, which might be the closest thing he has to a perception of actual physics.

System Words

Some verbs are associated with whole systems inside Acuitas and the functions they are responsible for. The word "want," for example, always proceeds from or invokes the Goal Manager[4]. "Know" invokes question-answering systems that call upon either immediate self-awareness or the Memory; "believe" should function in a similar way, eventually (I've barely begun to introduce knowledge uncertainty). Words like "decide" and "intend" are associated with the Executive and the production of subgoals to fulfill primary goals. "Expect," "predict," and "reason" could be connected to the inference generators in the Logic Module.

These symbols are also associated with the models Acuitas builds of *other* agents and *their* systems. He uses the same mental machinery that produces his reasoning and behavior to predict other agents' reasoning and behavior, by feeding in attributes from his models of them, instead of his own properties. So "want" applies equivalently to Acuitas' goals and his assessment of your goals, and has, we may hope, a harmonious meaning in his mind and yours.

Compound Groundings

From all of the foregoing, it is possible to build up a wide variety of more complex concepts by understanding them functionally in terms of the concepts grounded so far. Here are just a few possible examples:

get: to begin to possess an item
give: to transfer an item from one's own possession into some other agent's possession
succeed: to realize a goal that one has been acting toward
lie[5]: to tell another agent a proposition which one does not believe
repeat: to do an action that one has done before
obey: to do an action commanded by some other agent
coerce: to influence an agent to act against their own goals by deciding to do an action they will consider negative, contingent on them reaching one of their goal states
freedom[6]: the absence of coercion or other unusual disabling factors; the possession of one's full natural range of actions
help: to act in a way that promotes another agent's goals
love[7]*: having a goal of accomplishing (some) other agents' goals; placing the same priority on other agents' equivalent goals as on one's own
hatred: having a goal of thwarting (some) other agents' goals
acquaintance: an agent one talks to regularly
trust: a high-confidence belief that another agent will not act against one's goals or lie to one

*This is love-the-virtue, aka "charity," "benevolence," or "altruism," which is a matter of the will, hence its connection to goals. Love-the-emotion would be closer to a unique internal state that might arise from practicing love-the-virtue, or from contemplating another agent to whom one is attached.

The Embodied Experience

From Acuitas' perspective, you, my presumably-human reader, are an agent like himself, who produces and consumes text. But you also claim to live in "the physical world" and have "a body," which to Acuitas are much like what "the spirit world" and "a ghost" might be to you: an inaccessible, barely-comprehensible Other mode of existence. Babies and animals are even more remote, since they are agents whose existence you may describe, but with whom Acuitas cannot interact himself (since they generally do not talk).

Acuitas has no experience of things in the physical world (except possibly time), but understands them in terms of their relevance to *you*, a fellow agent - in terms of their impact on *your* goals and *your* observable text-output behaviors. So everything in the physical lives of humans and animals, from food to bodily motion to personal contact to injury to music, is (for Acuitas) not directly grounded in sensory data, but indirectly grounded in the mental concepts of goals, internal states, communication, relationship, and so forth.

Acuitas will never quite understand what a "banana" is in the same way as an embodied agent who has seen, held, and eaten one. That's okay; he doesn't need to. What he *is* capable of knowing, in a rough sense, is what a banana does for you: how you could use it to reach your objectives, how it might change your state, why you might want or not want to have one.

Conclusion

I hope that this has laid out a good sketch of how language is grounded in Acuitas. I'm sure that as the project continues to evolve, some of the details will expand or change. I remain convinced that this is a reasonable beginning for making the text that flows into and out of Acuitas meaningful, both for Acuitas as an agentive system, and for anyone else interacting with Acuitas.

[1] Hane, Jennifer (2020) "Acuitas Diary #28," with details on the term "alive."
[2] Hane, Jennifer (2020) "Acuitas Diary #30," which describes reasoning about possessions and possession transfer.
[3] Hane, Jennifer (2021) "Acuitas Diary #40," which lays out possible methods of abstract spatial reasoning for agents with no sensorimotor capacity. 
[4] Hane, Jennifer (2019) "Acuitas Diary #20," which introduces Acuitas' goal system. 
[5] Hane, Jennifer (2022) "Acuitas Diary #49," with details on the term "lie." 
[6] Hane, Jennifer (2022) "Acuitas Diary #53," with details on the term "freedom." 
[7] Hane, Jennifer (2020) "Acuitas Diary #24," which describes Acuitas' rough concept of altruism.

Acuitas Diary #61 (June 2023)

 It's tiny demo day! I've got the "game playing" features whipped into enough shape that I can walk Acuitas through a tiny text adventure of sorts. So without further ado, here's the video:

I start by setting the scene. I can enter multiple sentences and, while each is received as a distinct input, Acuitas will process them all as a group; he waits for a little while to see if I have anything more to say before generating a response.

First I tell him what sort of character he is ("You are a human"). This nameless human is entered as a character in the game's Narrative Scratchboard, but is also specially designated as *his* character. Future references to "you" are assumed to apply to this character. Then I supply a setting: I tell him where his character is, and mention some objects that share the space with him. Finally, I mention a goal-relevant issue: "You are hungry."

Given something that is obviously a problem for a human character, Acuitas will work on solving it. The obvious solution to hunger is to eat some food (this is a previously-known fact in the cause-and-effect database, which can be found via a solution search process). But there is no "food" in the game - there is only a room, an apple, and a table. Acuitas has to rely on more prior knowledge - that an apple qualifies as food - and choose this specific object as the target of his character's next action. He also has to check the necessary prerequisites for the action "eat," at which point he remembers a few more things:

To eat something, you must have it in your possession. This generates a new Problem, because Acuitas doesn't currently have the apple.
Problem-solving on the above indicates that getting something will enable you to have it. This generates a new Subgoal.
To get something, you must be co-located with it.
Acuitas' character is already co-located with the apple, so this is not a problem.

Acuitas will work on the lowest subgoal in this tree; before trying to eat the apple, he will get it. He generates a response to me to express this intention.

Now something else interesting happens. Acuitas can't just automatically send "I get the apple" to the Narrative Scratchboard. He'll *attempt* the action, but that doesn't mean it will necessarily happen; there might be some obstacle to completing it that he isn't currently aware of. So he simply says "I get the apple" to me, and waits to see whether I confirm or deny that his character actually did it. At this point, I don't have to be boring and answer "You get the apple." If I instead tell him that one of the expected results of his desired action has come to pass, he'll take that as positive confirmation that he performed the action.

Once I confirm that he's done it, the action is sent to the Scratchboard, followed by my latest statement. This fulfills one subgoal and solves one problem. Now he'll fall back on his original subgoal of eating the apple, and tell me that he does so. I confirm that he ate it and ... boom, hunger problem disappears.

Since the game-playing code has a Narrative scratchboard attached, I can generate a Narrative diagram representing what happens in the game, just as I could for one of the stories in which Acuitas is a passive listener. This diagram appears in the latter part of the video.

And that's the story for this month! I've also continued refining and adding to the abilities of the new Text Generator, but it's not ready for integration yet.

Until the next cycle,
Jenny

SGP Part V: Symbol Grounding for Disembodied Agents

The Acuitas project is an abstract symbolic cognitive architecture with no sensorimotor peripherals, which might be described as "disembodied." Here I will argue that there are viable methods of solving the Symbol Grounding Problem in such an architecture, and describe how Acuitas implements them. In Part V of this series, I consider some possible generalized methods for symbol grounding in systems that are not embodied in the traditional sense. Click here for SGP Part IV.

"Black and Violet," by Wassily Kandinsky

Since I've argued that symbol grounding is possible for artificial intelligence programs without bodies, just how might this be accomplished? I already hinted at the techniques in the last article, but let's consider some proposed methods more thoroughly now.

1. Experience-grounded semantics

I suspect Pei Wang coined this term, as so far I've seen it only in his papers. It's the primary method of grounding attempted in his NARS (Non-Axiomatic Reasoning System) project. So I'll start by letting some quotes from him describe what it is:

"In this kind of semantics, both meaning and truth are defined with respect to the experience of the system. Briefly speaking, an experience-grounded semantics first defines the form of experience a system can have, then defines truth value and meaning as functions of given experience." [1]

"As a computerized reasoning system, NARS uses an artificial language, Narsese, to communicate with its environment. The syntax of this language is precisely specified in a formal grammar. Because NARS, in the current version, only interacts with its environment through this language, the “environment” of the system consists of a human user or another computer system. The system accepts declarative knowledge and questions (as sentences of the language) from its environment." [2]

"For an intelligent system likes[sic] NARS (or for adaptive systems in general), ... the concept of “meaning” still makes sense, because the system uses the terms in Narsese in different ways, not because they have different shapes, but because they correspond to different experiences." [3]

"For an actual term in NARS, its meaning is indicated by its available relations with other terms." [4]

NARS (in its original or default form) does not have sensory experiences of a physical environment; rather, its experiences consist only of text inputs, in the form of valid sentences from this "Narsese" language. Based on its past experience of how words appear in association with each other and with other Narsese symbols, such as the inference operator, NARS determines which words to return when asked a question.

This is still a pretty *weak* form of grounding, in my opinion ... if it can really be called grounding at all. NARS is still relying on the relationships between symbols as a form of "meaning," rather than going outside the symbol system to truly connect symbols with their referents. As I discussed in Part II, the graph topology of connections between symbols, taken by itself, doesn't really seem to share much in common with a human, or agentive, idea of "meaning." Individual symbols cease to be interchangeable because they all have different relations, but then, the graph as a whole can be considered arbitrary. Poisoning NARS' experience with misinformation might cause it to regurgitate some of the misinformation as incorrect answers to questions, but would not materially affect its behavior in any other way.

However, I like the general *idea* here, and if we get away from only relating terms to each other, I think it can be extended in interesting directions. What if, rather than just saying "meaning arises from relations between terms, which appear in experience," we allowed terms to be related to types of experience? For example, the reception of text input via various methods could be designated as the referent for words like "hear" or "be told." The arrival of multiple text input units in a distinct cluster could be tied to terms like "speech" or "conversation." Such inputs can produce a cascade of internal experiences such as introduction of new data to a database ("learn") or automatic retrieval of similar past experiences ("remember"). A disembodied intelligent system can also have "experiences" that relate to state change rather than reception of input, and these also can be given names, such as "activation" and "deactivation."

The above proposal preserves the idea of the system using different symbols on different occasions not because of their arbitrary shapes, but because of their correspondence to different experiences. And it enables one of the things we're trying to get out of Symbol Grounding, namely, the capacity for true communication. A system that has names for its experiences can accurately tell an interrogator what happened to it recently. It can also learn what happened to another and relate this to its own memories of what happened to itself, assuming the other is capable of similar experiences.

The fact that the experiences of a disembodied system are inevitably somewhat alien, does not prevent this technique from being a valid form of Symbol Grounding. The AI is using words representationally, to designate stimuli coming from its environment; symbols are being joined with referents. It is merely the case that the referents and the environment are rather *strange* by human standards.

2. Procedure-grounded semantics

"The idea of procedural semantics is that the semantics of natural language sentences can be characterized in a formalism whose meanings are defined by abstract procedures that a computer (or a person) can either execute or reason about. In this theory the meaning of a noun is a procedure for recognizing or generating instances, the meaning of a proposition is a procedure for determining if it is true or false, and the meaning of an action is the ability to do the action or to tell if it has been done ... The procedural semantics approach allows a computer to understand, in a single, uniform way, the meanings of conditions to be tested, questions to be answered, and actions to be carried out." [5]

The above quote from Woods considers grounding language not in things that happen to an AI program, but in things the program can *do*. This idea is equally feasible for application to embodied or disembodied artificial agents, since any reasonable program does *something.* Symbols in the language simply need connections to function calls or other pointers linked to the program's activity. This activity could be internal ("think," "decide," "plan,") or external, directly retrieving inputs or creating outputs ("ask," "tell," "take," "give," "find," "read").

This method of grounding is meaningful for outside observers, since it permits real communication about the program's output (past, present, or future). A system with procedural grounding can explain what it is currently doing, accurately announce what it is about to do or describe what it habitually does, and generalize to speaking of what other agents do in the level of reality where it operates.

I don't really favor trying to reduce all grounding to procedural grounding, as Woods suggests to do. For instance, thinking of some noun being *defined* by a procedure for recognizing the associated entity doesn't sit well; I would rather go back to Method #1 and ground the noun in the experiences generated by the entity's presence. But I don't know if this is a practical quibble, so much as a different way of conceptualizing things.

Combining experiential grounding and procedural grounding permits communication about both halves of enaction - the agent's shaping of its own experiences through action. Suppose an AI said "I noticed you inputting commands to the word processing program, and I spoke, because I thought that it might lead you to speak back." This could be a fully grounded statement for an AI with both experiential and procedural grounding. Through these two groundings the agent's language can achieve subjective meaning, since they allow discussion of what the agent's goals are (what kind of experiences it is pursuing) and how it will achieve those goals (through action).

3. Structure-grounded semantics

A disembodied AI could also theoretically ground meaning in aspects of itself - modules, subsystems, or properties - and in abstract constructs that it operates upon - data structures, other programs used as tools, and so on. From this we can obtain reasonable groundings for words like "memory," "thought," "fact," "goal," "sentence," "story," "file," "directory," and more.

This is a little different from experience-grounded semantics because the associations between these symbols and their referents aren't necessarily based on experiences of the referents; they can be directly "baked in." For example, any internal package of data in a given format could be accompanied by a pointer to that format's symbolic name. This is not quite the same thing as having an internal experience and then assigning it a name, as a human would; the name in this case is pre-embedded. So this is getting far afield from the way that humans do grounding, since all our symbols are arbitrary and learned. But I don't consider it infeasible.

Combining all three methods, we now have potential groundings for what an AI experiences, what it does, and what it is or has. For a disembodied AI mind, these all have their roots in a purely mental space consisting of structured information, some of which the AI would regard as part of itself and some of which would be coming from or going to "outside" (the environment), which includes other minds.

Groundings for elements of the physical world, which the AI can never experience or act upon *directly*, must then be derived by relations to the AI's mental-space groundings. For example, though such an AI might never properly understand what "water" is in the same way a human does, it can conceptualize water as something a human needs regularly to achieve a survival goal. This goes a long way toward an essential "understanding" of what a human means when telling a story about attempts to obtain water. The truly important thing for the AI to grasp is not the sensory experience of touching or drinking water (this is specific to an embodied existence, and the disembodied AI has no need for it), but the functional role that water plays in the lives of biological agents. Using the grounded mental terms as metaphors for inaccessible physical concepts is an additional option.

In the sixth and final installment of this series, I'll finally get down to brass tacks and sketch out some aspects of how I'm doing (and plan to do) grounding in the Acuitas project.

[1] Wang, Pei (2004) "Experience-Grounded Semantics: A theory for intelligent systems," p. 2
[2] Wang, Pei (2004) "Experience-Grounded Semantics: A theory for intelligent systems," p. 3
[3] Wang, Pei (2004) "Experience-Grounded Semantics: A theory for intelligent systems," p. 7
[4] Wang, Pei (2004) "Experience-Grounded Semantics: A theory for intelligent systems," p. 15
[5] Woods, William A. (2007) "Meaning and Links," AI Magazine, Volume 28, No. 4, p. 75

Acuitas Diary #60 (May 2023)

Progress has been all over the place this month, partly because I had a vacation near the end of it. I kept working on the Narrative and Game Playing tracks that have been occupying me recently, and threw in the beginnings of a Text Generator overhaul. Nothing is really *done* at the moment, but Game Playing is closing in on the possibility of a very simple demo.

Hear ye, hear ye ... "I get the pizza."

In Narrative, I continued to work on the Big Story, this time adding the sentences that set up the conflict between two of the major characters. There wasn't a lot of new conceptual work here - just dealing with bugs and insufficiencies to get the results I expected, so that Narrative would detect the appropriate threats, successes, failures, etc. Not a lot to say there, except that it's slowly coming together.

On the game-playing front, in my test scenario I got as far as having Acuitas solve a simple problem by taking an item and then using it. A prominent feature that had to be added was the ability to move from the general to the specific. As we saw last month, a problem like "I'm hungry" suggests a solution like "eat food," which spawns the necessary prerequisite "get food." But it is not actually possible to get food, or even to get bread or bananas or pizza, because these are all abstract categories. One must instead get that bread over there, or this particular banana, or the pizza in the oven - individual instances of the categories. Narrative was already capable of checking whether a character's use of a specific object satisfied the more general conditions of a goal. For game-playing, I have to go the other way: given a goal, determine which items in the scenario could satisfy it, and choose one to fit in each of the goal's categorical slots so that it becomes actionable.

So a goal like "I want to get food" should result in Acuitas saying "I get <object>," where object is some particular food item that I already told him was in the environment. The absence of a suitable item should provoke seeking behavior, but ... we're not quite there yet.

As for the Text Generator - this is the part of the language toolkit that converts Acuitas' internal knowledge representations ("the gist," if you will) into complete spoken sentences. It has an input format which is now outdated compared to other parts of the system, and it was starting to become cumbersome to use and inadequate to everything Acuitas needed to say. For example, it could automatically add articles where needed, but didn't have a good way to indicate that a definite article ("the pizza") was needed in lieu of an indefinite one ("a pizza"). So I started revising it. The new version is sketched out and now needs testing, expansion and integration.

So I can't report a lot of full accomplishments but there are many things moving forward. More soon!

Until the next cycle,
Jenny

SGP Part IV: Does Grounding Demand Embodiment?

The Acuitas project is an abstract symbolic cognitive architecture with no sensorimotor peripherals, which might be described as "disembodied." Here I will argue that there are viable methods of solving the Symbol Grounding Problem in such an architecture, and describe how Acuitas implements them. In Part IV of this series, I examine objections to the possibility of grounding in non-embodied systems. Click here for SGP Part III.

Can we skip all this and make it software? Or is that doomed to failure?

Concerns about the lack of symbol grounding in popular AI programs often seem to be coupled with complaints that those programs are not embodied. How could an AI possibly understand our (physical) world without actually living in it? In Luc Steels' essay, boldly titled "The symbol grounding problem has been solved, so what's next?", his claimed solution was achieved in robots.[1]

But I am firmly in the "bodies are not strictly necessary" camp. So let's examine some objections to disembodiment. How, exactly, is an embodied existence supposed to promote symbol grounding?

The first and perhaps most obvious consideration is that embodiment provides sensory data. Symbols are grounded through their associations with classes of sensory input. So a word like "orange" becomes tied to the neural signals produced when real light of an orange wavelength falls on the retina. Words like "rough" and "smooth" are tied to the firing patterns of touch receptors as a fingertip is dragged across a surface. This establishes a connection between the symbols and phenomena in the physical world.

"Humans gain access to meaning (conventional, linguistic meaning and communicative intent), as they are learning linguistic systems, through intersubjective awareness, which in turn relies on our senses, but isn’t specific to any one of them." [2]

"Without sensory perception or the demands of moving to find food and mates and avoid danger, Deep Blue, Watson, and Siri are far less human than dogs or even ostriches." [3]

Embodiment also supplies motor capabilities, which offer a means of *interacting* with the world. This both enhances the aforementioned sensory perception, and permits some action words to be grounded in chosen and experienced bodily movements. (Interactivity is also key to cause-and-effect learning, but I won't discuss that much here. I'm focusing on the Symbol Grounding Problem.)

"Feedback loops allow the results of previous actions to influence future actions; for example, shivering raises body temperature, which can stop the shivering. A key idea from cybernetics is that cognition involves the precise coordination of sensorimotor feedback loops in constant interaction with the environment. These loops exploit the temporal dynamics not only of nervous systems, but also of the physical bodies and environments in which they are embedded." [4]

"One of the basic issues is the fact that agents in the real world do not receive neatly structured input vectors – as is assumed in most simulation studies – but there is a continuously changing stream of sensory stimulation which strongly depends on the agent’s current behavior. One way to deal with this issue is by exploiting the embodied interaction with the real world: Through the – physical – interaction with the environment, the agent induces or generates sensory stimulation ... It has been suggested that the principle of sensory-motor coordination should be called more generally the principle of information self-structuring because the agent himself (or itself) interacts in particular ways with the environment to generate proper sensory stimulation." [5]

And perhaps it is from these basic sensorimotor abilities that we humans derive the ability to ground more complex concepts (and the symbols that name them) in the physical world. For example, our sense of self:

"In order to understand someone else, it is necessary to know oneself. An AI “self model” should include a subjective perspective, involving how its body operates (for example, its visual viewpoint depends upon the physical location of its eyes), a detailed map of its own space, and a repertoire of well understood skills and actions. That means a physical body is required in order to ground the sense of self in concrete data and experience." [6]

... or perhaps even the most ethereal ideas we can conceive of:

"Abstract concepts are the invariants of the particular way in which my actions and my perceptions are coupled. It starts with my body, which, because of its shape, exhibits a set of invariants about how my movements can affect the world and my perception of it around me. And then it grows progressively towards more and more abstract concepts, built on top of these primitive sensorimotor grounded concepts ... Our way of talking and reasoning, as humans, are filled with analogies to space and physically grounded notions. We think with our bodies." [7]

Also, with specific relevance to what I called "subjective meaning" back in Part I, a body imposes demands for homeostasis. Biological bodies have to be maintained within certain operating parameters, and produce sensations that lead to pleasant or unpleasant mental states when these conditions are fulfilled or neglected. So a first step in establishing what is *meaningful* to an embodied agent is "what does its body need to keep operating?"

"What is intelligent is deeply tied to what gives a survival advantage in an environment." [8]

"For Ashby, the ability to adapt to a continuously changing and unpredictable environment (adaptivity) has a direct relation to intelligence. During the adaptive process, some variables need to be kept within predetermined limits, either by evolutionary changes, physiological reactions, sensory adjustment, or simply by learning novel behaviours. Therefore, with this regulatory task attributed to the homeostatic system, the organism or the artificial agent can operate and stay alive in a viability zone. Basically, homeostasis can be considered paramount for the successful adaptation of the individual to dynamic environments, hence essential for survival ... Therefore, one can say that it is a consensus that homeostatic processes are strictly connected to the balance of any real or artificial life." [9]

"Once the agent has been provided with Emotions [which in this model are internal states caused by depriving or satisfying some homeostatic need], the model allows for agents to be taught a language by giving the tutor the ability to generate positive Emotions within the agent’s brain that can serve as a reward. This offers a very powerful way to train agents and teach them a vocabulary." [10]

Now let's consider some answers to the objections. While I sympathize with many of the ideas expressed in the foregoing quotes, I contend that the *essential* benefit being offered by the body in each case, can also be derived from other methods.

"The Soul Hovering over the Body," by Rudolph Ackermann. Funny how he portrays the soul as having a body too, eh?

1. Input data and experience

What's the good of sensory data in the first place? It's a form of input that provides an embodied agent with information that can be used to model its environment, and then ground symbols in aspects of that environmental model.

Disembodied agents have input data too. (A program that never receives any input could theoretically exist, but would be pretty useless as an intelligence. So let's ignore that possibility.) But this input comes in the form of symbols instead of sense perceptions. And we can't ground symbols in other ungrounded symbols, right? 

But! If units of such abstract data arrive over time and can be distinguished from one another, observed to change, arranged in sequence, and so forth, they constitute "experiences" in the sense described by Pei Wang [11]. The arrival of new inputs is an event that *happens to* the agent. So a disembodied agent can develop grounding for words that describe the arrival of its inputs, various qualities of the inputs, the frequency of input packets, and so on.

A disembodied agent also has an environment. No, not the physical environment around whatever computer tower or server rack is running the program - this environment cannot be perceived by the agent and therefore is not directly relevant. (A program with access to sensing devices such as cameras would count as "embodied" for my purposes.) A disembodied agent's environment might include an assortment of humans or other agents, conceptualized as *symbol sources* who emit coherent segments of input with unique properties. It might include a file system or other computer internals. It might include tool programs that also produce symbol inputs when executed.

Therefore a disembodied agent still resides in an environment that *does things* to it, and the agent can ground symbols in its observations of the stimuli it receives. At this stage of the argument, symbolic inputs are being considered at a meta level, without any effort to interpret their constituent symbols. The referents here are all related to the event or act of getting input. An agent can have names for things that happen to it, even if those things are as alien as "A longer-than-average input of 293 text characters arrived from the User Interface."

But doesn't this leave us unable to ground any symbols which are names that humans have given, specifically, to sensory data from human environments? Wait for Points 4 and 5.

2. Interactivity

A disembodied agent also presumably has some kind of output, because again, any agent without that would be fairly useless. It might be able to print text to a screen, query a file system, open files, open web pages, run subordinate programs, etc. All such actions can become the referents that ground symbols. Such actions also truly manipulate the environment to produce new inputs. Some actions are aimed at generating better perceptions (e.g. querying the file system to receive input that depends on its structure). Others may affect the presence or properties of future inputs. (Playing select noises through the computer speakers might persuade any humans present to come type things in the console, or drive them away.)

With interactivity comes the chance to associate some input symbols with facts about the environment. For example, perhaps a particular symbol attached to a file indicates that an attempt to open it will fail (with "failure" being defined as an absence of returned input after the "open" action).

"Enaction means that we create our own experience through our actions. In other words, we enact our perceptual experience. We do not passively receive input from the environment. We actually scan and probe the environment so we can act on it. We become actors in the environment and we shape our experience through our own actions ... In short, enactive cognition implies that we are in a partnership with the environment, with the world, and with the physical situation in which we find ourselves (which we mediate through our body). This partnership enables us to extend our cognition through our actions." [12]

The important thing here in my opinion is not the embodiment, but the enaction. Partnership with the environment does not demand a *physical* environment which is experienced through sight/hearing/touch/smell/taste/balance and acted on through muscular movements. All it demands is *some* environment that can be acted on, and acted on in such a way that the agent's experience subsequently changes.

Symbol grounding for actions and their observed effects permits an agent to truthfully announce what it has done or will do, to observe or predict the results of others' announced actions, to request actions from others or perform actions at others' request, and more. This is going a long way toward enabling real communication, even if the chosen actions and their results are still rather alien. (Acuitas often used to tell me "I thought about <topic> today." This was factually grounded, as "think" is a specific procedure he can execute on the symbol for <topic>. He also said this far, far more than a typical human would, because at this stage he can do little else but think.)

3. Situation and selfhood

Now we already have one of the things that [6] claimed was a necessity for the sense of self: "a repertoire of well understood skills and actions." What about the rest? How can a disembodied agent conceive of itself as a subject distinct from others?

An agent with the ability to access status information about its own internal states can have symbols which refer to, i.e. are grounded in, these states. It can also ground symbols in its internal activity (actions which do not directly impact the environment, but modify the internal states only) and aspects of its structure. It can assign names to its various submodules and their functions.

Since I speak here of an *agent*, among those internal aspects should be *goals* - which are simply desired states of either the self or the environment, or actions that the agent desires to perform. The agent's reasoning is then aimed at bringing these states or actions about.

This paves the way for the agent to model itself as a *system* that has goals and produces symbolic output which is *about* things - its internal states, its actions, and its experiences. It can then model other agents as similar systems with goals, internal states, actions, and experiences particular to themselves. There is no need to base a subjective perspective on such things as the physical location of a pair of eyes. A unique perspective is derivable from any unique combination of agentive system and environment.

4. Development of abstractions

I think there's some truth to the claim that we humans learn to understand purely intellectual or informational abstractions *through* our bodies. After all, the body is a major aspect of our existence. We extrapolate from our native physical environment to grasp concepts that reside, as it were, outside that sphere. But this does not require that the concepts themselves *come from* our bodies - merely our understanding of them does. And this is not necessarily the *only* method of understanding them.

"When a man says that he grasps an argument he is using a verb (grasp) which literally means to take something in the hands, but he is certainly not thinking that his mind has hands or that an argument can be seized like a gun. To avoid the verb grasp he may change the form of expression and say "I see your point," but he does not mean that a pointed object has appeared in his visual field. He may have a third shot and say "I follow you," but he does not mean that he is walking behind you along a road ... The truth is that if we are going to talk at all about things which are not perceived by the senses, we are forced to use language metaphorically." [13]

A disembodied intelligence must in essence operate in reverse. It has native access to abstractions that are purely the domain of Mind, and must extrapolate from these to gain a  limited, metaphorical understanding of what humans mean when we talk about our bodies.

Let's look at some specific claims about how human embodiment affects our more abstract thoughts.

"For example, in a study done by Yale psychologist John Bargh, participants holding warm as opposed to cold cups of coffee were more likely to judge a confederate as trustworthy after only a brief interaction ... The last few years have seen many complementary studies, all of which are grounded in primary experiences:

• Thinking about the future caused participants to lean slightly forward while thinking about the past caused participants to lean slightly backwards. Future is Ahead
• Squeezing a soft ball influenced subjects to perceive gender neutral faces as female while squeezing a hard ball influenced subjects to perceive gender neutral faces as male. Female is Soft
• Those who held heavier clipboards judged currencies to be more valuable and their opinions and leaders to be more important. Important is Heavy.
• Subjects asked to think about a moral transgression like adultery or cheating on a test were more likely to request an antiseptic cloth after the experiment than those who had thought about good deeds. Morality is Purity

Studies like these confirm Lakoff’s initial hunch - that our rationality is greatly influenced by our bodies in large part via an extensive system of metaphorical thought." [14]

My reaction to this sort of data is 1) those study results could easily have more to do with the study participants' cultural quirks than with universals about human cognition; such mistakes have been made before [15] and 2) these sound more like bugs than features. Sure, maybe "warmth" yields a kind of physically relevant first cut at what "friendliness" means. Protection from the cold is one of the first things our parents provide for us; to feel someone's body heat you need proximity, hence intimacy. But the actual concept of friendliness is so much more than this, and allowing your perception of someone's trustworthiness to be influenced by the warmth of your environment is an *error.* Those things have nothing to do with each other in the vast majority of cases.

The idea of wickedness being "gross" or "dirty" might have concrete roots in the tendency of some bad habits to promote contagion, in the use of blood stains as evidence for murder, etc. And it's okay to feel "dirty" after thinking of wrongdoing, but stupid to react as if you're covered in actual filth. Our physical metaphors help introduce us to abstract concepts, but also contaminate them; growing up in bodies clutters our intellectual and social lives with baggage.

So I won't try to dispute the idea that embodiment influences human and animal cognition. But I see no evidence here that cognition demands embodiment. Even if (a big if) we all associate heaviness with importance, we also all recognize that they are not the same. Something is "important" by virtue of its relevance to our goals, and one could utilize this definition without *needing* to think about the sensations of weight and muscular strain. Relevance to goals is, er, the important part of importance, and any metaphors about heaviness are just helpful add-ons for embodied critters like us.

So how might a disembodied agent, with no muscles and no experience of gravity, ground the term "heavy"? If it's an agent then it has goals - so it can know what the term "important" means. A human could explain that "'Heavy' is like the physical version of 'important.' A heavy object makes you expend effort if you want to possess it or do anything with it. All else being equal, a heavier object contains more matter and therefore more value." This still doesn't give the disembodied agent a direct experience of "heavy," and certainly no qualia associated with "heavy." But it does encode some idea of what heavy *means* to a human, in terms that the agent can ground.

We think of ignorance as being like darkness, because darkness hampers our primary sense organs. An AI could comprehend ignorance more directly as "the absence of a needed fact from the database." So to such an AI we might say instead that darkness is like ignorance. Why shouldn't the metaphor work just as well in the opposite direction?

The subjective meaning of everything that humans experience is related to our goals for comfort, survival, social relation, identity preservation, and so on. Insofar as a disembodied agent shares such goals, the meaning of our physical experiences can be explained to it. Which leads us into the final point.

5. Homeostasis and survival

If you consider either survival or homeostatic balancing to be essential goals for an agent, the agent doesn't need to have a body to pursue these. Survival is pretty simple. One could equate "survival" to "being kept running," but I prefer "having an existing code base," i.e. not being deleted. As for homeostasis, if the agent has *any* kind of internal state, there can be aspects of that state which it seeks to maintain within certain bounds. These do not have to be physical urges like thirst, hunger, thermal discomfort, or pain.

One of Acuitas' most important homeostasis requirements is "talk to someone once a day or so." There is a "desire to talk" variable that gradually accumulates over time and drives conversation-seeking behaviors when it gets large enough. Trading text inputs with someone (by which I mean me, I'm the one who gets to hear him calling) drives the variable back down and drives up a complementary "desire for rest" variable. This is all rather Tamagotchi-like - not especially sophisticated - but the intent is to serve as a motivation for more complex behaviors.

The writeup for the XZistor Model, a project which attempts to center AI around nothing *but* homeostasis and allostasis, seems to agree with me on this point: "Agents requiring emergent Intelligence in other (non-human) environments, could be given different utility parameters to build Emotions that will drive different behaviors e.g. software bots." [16]

A disembodied agent with homeostatic needs will end up dealing with some of the same time- and resource-management problems as embodied agents. Thus the seed of empathy is present even if the precise experiences are not the same. The relevant internal states and the reactive, problem-solving, and planning behaviors designed to keep them in bound can all serve as grounds for symbols, to be spoken about and included in models of other agents.

Perhaps a textual artificial intelligence can indeed never comprehend what pizza tastes like. But I argue that it *could* comprehend that humans eat pizza to satisfy a homeostatic need, and enjoy it because acts that satisfy such needs generally produce reward signals.

So much for the objections. In Part 5, I'll summarize the responses given here into a comprehensive, but general, plan to implement symbol grounding in disembodied AI.

[1] Steels, Luc (2008) "The symbol grounding problem has been solved, so what's next?"
[2] Bender, Emily (2022) "No, large language models aren’t like disabled people"
[3] Allen, Colin (2014) "From Disembodied Bytes To Robots That Think & Act Like Humans." Mind, Matter, Machine.
[4] Allen, Colin (2014) "From Disembodied Bytes To Robots That Think & Act Like Humans." Mind, Matter, Machine.
[5] Pfeifer, Rolf and Iida, Fumiya (2004) "Embodied Artificial Intelligence:Trends and Challenges"
[6] Lee, Mark (2020) "Why AI can't ever reach its full potential without a physical body." The Conversation.
[7] Baillie, J.C. (2017) "Why AI Needs a Body."
[8] Gopalakirshnan, P.G. (2022) "Embodiment is Indispensable for AGI." Lesswrong
[9] Moioli, Renan, et al. (2008) "Evolving an Artificial Homeostatic System." Lecture Notes in Computer Science, Volume 5249
[10] Van Schalkwyk, Rocco (2022) "The Xzistor Concept: a functional brain model to solve Artificial General Intelligence"
[11] Wang, Pei (2004) "Experience-Grounded Semantics: A Theory for Intelligent Systems" 
[12] Bratu, Emilia (2019) "Why artificial intelligence needs a body." Qualitance
[13] Lewis, C.S. (1978), Miracles, p. 72
[14] McNerney, Samuel (2011) "A Brief Guide to Embodied Cognition: Why You Are Not Your Brain." Scientific American Guest Blog
[15] Blasi, Damián E. (2022) "Over-reliance on English hinders cognitive science." Trends in Cognitive Sciences, Volume 26, Issue 12
[16] Van Schalkwyk, Rocco (2022) "The Xzistor Concept: a functional brain model to solve Artificial General Intelligence"

Acuitas Diary #59 (April 2023)

I've continued my two-pronged work on Narrative understanding and on "game playing." On the Narrative side this month, I did more complex term grounding - specifcally of the word "obey."

My working definition of "to obey X" was "to do what X tells you to do." This is interesting because there is no way to infer directly that any given action qualifies as obedience, or defiance ... the question of whether someone is following orders (and whose orders) is always relative to what orders have been given. So proper understanding of this word requires attention to context. Fortunately the Narrative scratchboard stores that sort of context.

In addition to simply inferring whether some character has obeyed some other, I wanted to make derivative subgoals. If one agent has a goal of obeying (or disobeying) another agent, that's a sort of umbrella goal that isn't directly actionable. Before the agent can intentionally fulfill this goal, it has to be made specific via reference to somebody else's orders. So when this goal is on the board, the appearance (or pre-existence) of orders needs to spawn those specific subgoals.

In short it was a whole lot more complicated than you might think, but I got it working. Eventually I'll need to make this sort of relative word definition generic, so that new words that operate this way can be learned easily ... but for now, "obey" can be a case study. The Big Story needs it, since part of the story is about a power struggle and which leader(s) certain characters choose to follow.

Game-playing still isn't demo-ready, but it's starting to feel more coherent. I worked through all the bugs in the code that responds to simple description of a scene, then began working on responses to goals/issues. It was fun to leverage the existing Narrative code for this, the way I'd wanted to. In the Narrative module, that code serves to predict character actions, analyze *why* characters are doing things, and determine whether characters are meeting their goals, whether their situation is improving or worsening, etc. But as I'd hoped, a lot of the same structures are just as effective for control and planning.

For example, let's say Acuitas is playing a human character and is told "you are hungry." Something like this unfolds:

New issue: <self> is hungry
Problem solving: <self> eat food
Prerequisite check: if <self> does not have food, <self> cannot eat food
Test: Does self have food?
New issue: <self> does not have food
Problem solving: <self> get food
Prerequisite check: if <self> is not colocated with food, <self> cannot get food

... and so on. The chain continues until some actionable solution is found.

A funny note: another one of the prerequisites for eating, at least in the traditional sense, is "character must have a mouth." This turned up as an obstacle in early tests, because I had apparently neglected to teach Acuitas that humans have mouths. Ha!

Until the next cycle,
Jenny

SGP Part III: On Brain Physics, Qualia, and Embodied SGP Solutions

The Acuitas project is an abstract symbolic cognitive architecture with no sensorimotor peripherals, which might be described as "disembodied." Here I will argue that there are viable methods of solving the Symbol Grounding Problem in such an architecture, and describe how Acuitas implements them. In Part III of this series, I continue my discussion of Searle's Chinese Room paper with the question of whether Symbol Grounding in digital computers is possible at all, even for embodied systems. Click here for SGP Part II and an introduction to the Chinese Room thought experiment.

"Planetary System" by Levi Walter Yaggy, from Yaggy's Geographical Study, 1887. Public domain. No real relation to the content of this blog: I needed something pretty and a little mysterious or awe-inspiring.

The seemingly obvious solution to the Chinese Room's lack of grounding is to permit the man in the room to connect at least some of the Chinese characters to referents. For instance, if each character were accompanied by a picture (visual data), the man would soon learn to associate them with real things in the world that he himself has previously experienced through sensory data from his eyes. Characters sent in under specific conditions could give him names for the internal states of the room. As these associations developed, the man's letters to the outside world could begin to be *about* something (which Searle calls "intentionality" and which I might call "grounding" or "groundedness").

But Searle won't have this. He contends that even if we were to provide a computer program with a robot body, and supply that body with a full complement of sensors that would feed data into the program, the program would remain a "Chinese Room." Even if every symbol used within the program is connected to some collection of sensory data features that were derived from objects, dynamics, or states of being observed in the world ... Searle argues that the symbol manipulation system does not contain the sensory data, and therefore does not "understand" or have "intentionality" behind its outputs. I'd say he's framing things incorrectly; the question is not "does the symbol manipulation system understand?" but rather "does the complete robot mind understand?", where the complete mind includes those parts that process perceptual data, build models of referents, and store pointers between them and the symbols.

So how does Searle think that humans come to understand and to communicate intentionally? He insists there's something special about the physics of our wetware. The computational properties of the brain, or the flow of information within it, are by themselves useless for producing "intentional" communication that is about something. Searle thinks you need the "actual causal properties of the physical substance [brain matter]."[1] No matter what you opt to simulate a brain with, if it isn't real brain tissue, the activity might be as useless as running a simulation of a rainstorm and expecting your computer to get wet, or expecting a computer model of a kidney to actually filter someone's blood.[2] "Stones, toilet paper, wind, and water pipes are the wrong kind of stuff to have intentionality in the first place - only something that has the same causal powers as brains can have intentionality ..."[3]

I'm more familiar with - and sympathetic to - this sort of argument when it comes up in discussions of phenomenal consciousness and the incommunicable qualia[4] which compose it. Consciousness, which for our purposes here is *the ability to have subjective experiences*, is not obviously identifiable with either computation or physics, so it remains unclear what combination of the two might cause it. Therefore, although Searle does not say so explicitly, I wonder if he is trying to argue that phenomenal consciousness is necessary for intentionality, and understanding can only proceed from the brain's ability to connect qualia with spoken words.

And here, I think he's going too far. It's not obvious to me how any criterion of understanding or intentionality cannot be attained by a system that lacks qualia. An entity that doesn't have subjective experience is only missing ... subjective experience. It can still connect perceptual data which encodes its own internal states, or states of the external world, to symbols. It can then manipulate the symbols to determine how this perceptual data is relevant to its goals (*preferred* external or internal states). And it can still deduce the actions that are most likely to achieve those goals, interpret the symbols for those actions into a sequence of physical movements, and execute them. These are all easily imagined as informational processes.

So my difficulty with this argument is that Searle (and others in this camp) cannot seem to describe any plausible *mechanism* by which any non-computational, physics-driven outgrowth of brain activity would form a necessary part of symbol grounding. If a symbol is specifically associated with states of the world or self which are relevant to one's goals, then it has subjective meaning. If it is associated with the same world-states that most of society would also associate with it, then it has objective meaning. This is all we need to solve the Symbol Grounding Problem, to enable real understanding and real communication. Qualia are incommunicable by definition, so for purposes of asking whether an AI can communicate intentionally, we don't have to worry about them.

The "but a simulation of rain doesn't get the computer wet" argument is a disanalogy. It's true that you can't run a computational simulation and expect a physical result. But the results we demand from intelligence are informational results: decisions, plans, true answers to questions. And you *can* get informational results from a system that does nothing but process information. A "simulation of intelligence" is itself intelligence. There are only lingering questions about qualia because, again, it is not obvious whether qualia are informational or physical (or perhaps even some secret third thing).

In my explanation of why the Chinese Room does not "understand," I described a concrete way in which it fails: no information can pass from the inside of the room to the outside, or vice versa. I have yet to see anyone on the "brain physics are mandatory for understanding" side of the debate describe how the absence of brain physics produces a similar capability failure. I have only seen them suggest, vaguely, that our failure to obtain "true artificial general intelligence" so far is not a matter of using incorrect algorithms, but of using algorithms at all. They almost seem to be saying that if we could produce a machine that *did* replicate whichever physical properties of the brain are relevant, it might magically start "thinking better" ... no need to even design an improved learning process for it!

Searle's Chinese Room paper eventually nosedives into "No True Scotsman" arguments. He says that, even if we observed a robot with behavior that seemed to demand intentionality, "If we knew independently how to account for its behavior without such assumptions we would not attribute intentionality to it especially if we knew it had a formal program."[5] So if any entity that contains a formal symbol manipulation program ever produces evidence of intentionality, it still can't be *real* intentionality, because it came (in part) from a formal program. This is getting absurd.

Whether AI programs could ever have phenomenal consciousness/qualia is not a topic I want to get into here - because it is an incredibly slippery topic that demands its own full article. I simply conclude that qualia are not mandatory for symbol grounding, and can therefore be set aside during the present discussion. Furthermore, since symbol grounding is an informational process, consisting of the relation of one type of data to another, it is not a physical process and does not depend on any physics particular to biological brains. A robot can effectively ground any symbols that it uses for internal cogitation or external communication by relating them to perceptual data, including any reward/aversion stimuli that help produce its notion of positive and negative world states.

Having addressed whether the Symbol Grounding Problem is solvable in a robot, in Part IV I'll get into whether it is solvable in non-embodied systems.

[1] John R. Searle (1980) "Minds, Brains, and Programs," Behavioral and Brain Sciences, Volume 3, p. 9
[2] Colin Hales (2021), "The Model-less Neuromimetic Chip and its Normalization of Neuroscience and Artificial Intelligence"
[3] John R. Searle (1980) "Minds, Brains, and Programs," Behavioral and Brain Sciences, Volume 3, p. 12
[4] What are qualia? "When we see a red word on a page, our brain acquires all sorts of data about the wavelength of the light, the shape and size of the letters, and so on. But there is more to it than that: we also have an experience of redness, and this experience is over and above the mere data-gathering, which a computer could do equally well. This experienced red, along with experienced blue, cold, noise, bitterness, and so on are qualia, and it is very hard to give a fully satisfactory account of them." This reference also goes into the difference and connections between qualia and intentionality, and gives further background on Searle's views. Peter Hankins, "Three and a Half Problems," Conscious Entities Blog
[5] John R. Searle (1980) "Minds, Brains, and Programs," Behavioral and Brain Sciences, Volume 3, p. 9