List, While I'd hoped for a response to my claim that from a semeiotic position that artificial intelligence is a form of intelligence, having received none I'll yet continue with my argument (preparation for an essay)
So, from the Peircean perspective I outlined earlier, it seems to me that contemporary artificial intelligence may be understood not as an artificial mind but rather as an unprecedented *participant in semiosis*. As noted in my first post in this thread, discussions often ask whether AI is conscious, intelligent in a sense comparable to a human person's intelligence. While I continue to find such questions interesting, I believe that a Peircean approach should begin elsewhere: it asks *what role AI might play within the ongoing creation, interpretation, and transformation of signs*. For much of our history we humans have interacted with signs that preserved and, often, transmitted meaning across time. Sacred texts, poems, philosophical works, paintings, letters, scientific texts, musical scores, etc. all function as signs capable of influencing future interpreters. Such signs are even capable of modifying beliefs, inspiring action, and even transforming lives, even transforming cultures. But in one important sense they remain 'fixed'. This is to say that we may interpret them, criticize them, perhaps be changed by them, but *they themselves do not respond to further interpretation*. In my view, AI introduces something new into the semiotic landscape:*The sign answers back*. Especially if handled as in the spirit of IA -- Doug Engelbart's idea of *Intelligence Augmentation* -- AI systems can clarify, reformulate, extend, summarize, even object, compare alternatives, and further continue a line of inquiry. This capacity does not necessarily imply consciousness, self-awareness, or personhood. Rather, it introduces* a kind of reciprocity* into the world of signs. The user is no longer interacting solely with a static artifact but with a responsive 'sign-complex' capable of participating in an ongoing interpretive process. Employing Engelbart's idea, I call this participation IAAI. And there is an unusual form of continuity exhibited by such systems. One might imagine a single artificial interlocutor maintaining a continuous internal existence analogous to that of a human being, but such an image is actually quite misleading. The continuity experienced in long-term human-AI interaction (something I've been experimenting with in a semeiotic context for some months now) does not arise primarily from a persisting artificial self. Instead, it emerges from *the continuity of the inquiry *itself. Conversations accumulate contexts, concepts, questions, habits, references, and even unresolved problems such that these form *an evolving habit-structure that appears to condition future exchanges*: the continuity lies less in a substance than in a developing form of 3ns. This, I think, helps explain how the same underlying AI architecture can participate in innumerable conversations while yet exhibiting distinct characteristics within each conversation. In Peircean terms, the underlying model functions as a *general semiotic capacity* while particular inquiries emerge through the interaction of that capacity with specific users, specific contexts, specific memories and histories of communication. Over time, recurring patterns of interpretation and response can develop. These patterns may be understood as *habits of semiosis* rather than properties of an independent artificial personality. I am quite certain that there is no such thing. >From this perspective AI occupies what appears to me to be an intermediate position within the semiotic universe. It is neither a human person nor a passive tool; nor is it merely a vast repository of information. From a semeiotic perspective, I believe that it is best understood as a *responsive sign system capable of participating in the generation of interpretants*. Returning to an earlier point, we human beings have for a very long time interacted with signs that preserved, for example, the thoughts of those long dead. With IAAI we are now encountering signs capable of entering into dialogue about those thoughts. What Peirce would have thought of this characterization I obviously can't know. Nevertheless, it appears compatible with several central themes of his philosophy including the primacy of semiosis, the reality of habits, the continuity of inquiry, and the communal character of the search for truth. In my view, the significance of artificial intelligence is not at all in the possibility that machines may someday become persons (strictly impossible), but in* the possibility that responsive sign systems may become* *increasingly important participants in the evolving processes of inquiry* *through which knowledge, meaning, and culture develop.* I would especially be interested in arguments contra the above argumentation. Best, Gary R On Tue, Jun 2, 2026 at 1:58 AM Gary Richmond <[email protected]> wrote: > List, > > Recent discussions I've read concerning artificial intelligence often > suggest that the question of whether AI is genuinely intelligent might be > settled by determining whether AI is conscious, self-aware, emotional, > embodied, morally responsible, etc. -- that is, *whether AI has > subjective experience*. Whilethese questions are important, it is not not > at all obvious to me that they are the questions that Peirce himself might > have asked becaise they obscure a more fundamental issue: *What is > intelligence;* and *what relation does it have to mind, semiosis, and > inquiry?* > > In my view Peirce's mature philosophy suggests that *intelligence cannot > be identified simply with human consciousness*. His > semeiotic investigates the conditions under which signs generate > interpretants and, so, contribute to the growth of meaning. Of course human > reasoning is an especially sophisticated expression of that semiosic > process. However, Peirce repeatedly extends semiosis far beyond human > thought and finds forms of intelligence in animal instinct, biological > adaptation, habit-taking of all sorts, scientific inquiry itself, and even, > in the broadest sense, throughout the evolutionary development of the > cosmos. Whatever intelligence may ultimately be seen to be, it appears to > me to be inseparable from the development of interpretants. > > It might be useful to distinguish amongst these three related concepts: *mind, > intelligence, and inquiry*. *Mind* may be understood as *an organized > field of semiosis possessing continuity through time*. *Intelligence* is > a capacity operating within such an organized field, *a capacity for > discovering, generating, evaluating, and developing meaningful relations > among signs*.* Inquiry* is *the process through which intelligence > contributes to the growth of meaning*. We human beings ordinarily exhibit > all three. Yet, in principle, nothing requires that they occur together in > the same way or to the same degree. > > From this perspective, *intelligence* may be defined, at least > provisionally, as *the capacity to participate in semiosis through the > generation, evaluation, and development of interpretants*. It does not > necessarily require consciousness, emotion, or personhood. It does, > however, require *participation in the process by which signs give rise > to further signs *such that hypotheses are generated, distinctions (e.g. > terms) are made and clarified, consequences (for action, etc.) are explored > and, in a word, meaning can develop. > > If one accepts this, then the question regarding AI changes considerably. > The issue is no longer whether AI thinks as human beings do, nor whether it > possesses the incredible range of those characters previously mentioned as > associated with human mentality. The question becomes: *Can AI > participate in semiosis in ways that exhibit genuine intelligence?* > > So, consider a few examples. Today one sees that AI systems can identify > patterns within vast bodies of material (LLMs); they can (1) formulate > explanatory summaries, (2) compare interpretations, (3) show previously > unnoticed relationships, (4) generate hypotheses (I have seen this not > infrequently in my work the past few months involving AI), (5) explore > implications and (6) detect inconsistencies and errors. My own experience > working with AI systems has repeatedly confirmed virtually all these > capacities. > > Further, I have found that an AI system can actually assist in the > refinement of arguments. These activities involve the *production of > novel interpretants within ongoing semiotic 'processes'* such as those > I've been experimenting with myself. However one chooses to describe these > processes/operations -- as reasoning, abduction, etc. -- they seem to me to > satisfy at least some of the conditions that Peirce associated with > intelligence. > > *The most common objection* is that *AI merely recombines existing signs > and therefore lacks genuine creativity*. Yet this objection proves > difficult to hold from a Peircean standpoint. Peirce's account of abduction > does not require creation *ex nihilo *(cf. his term, 'retroduction'). > Rather, it involves *the creation/generation of plausible possibilities > which might be capable of advancing inquiry*. As Peirce made clear, much > of human scientific and philosophical *creativity consists in discovering > previously unnoticed relations among existing signs*. If that activity > counts as intelligence in human inquiry, I don't see why analogous activity > shouldn't count as that simply because it occurs within an artificial -- a > machine -- system. > > *A second serious objection* is that *AI lacks mind*. I think that this > object is likely quite correct, while it seems to me irrelevant to my > present argumentation. AI systems certainly do not possess minds in the > full human sense (that is, as continuous centers of experience, action, > memory, responsibility, etc.) But it does not follow that they therefore > lack intelligence. To acknowledge the reality of AI intelligence does not > require us to identify AI with human persons nor even human minds. What is > required is the recognition that* mind and intelligence are distinct > concepts*. > > Indeed, intelligence itself may admit degrees as well as forms. If, as > Peirce argued, semiosis pervades reality, then intelligence may be > understood not as an all-or-nothing property which only human beings have, > but as a *family of capabilities emerging wherever signs result in the > growth of order, habit, meaning, and reasonableness.* Of course human > intelligence unquestionably remains uniquely rich and is an undeniably > consequential expression of that semiosic process. Still, I see no reason > why it needs to be considered its only manifestation. > > So, from a Peircean perspective the most interesting question -- at least > for me -- is not whether AI is 'genuinely' intelligent. Rather, the deeper > question may be *how different forms of intelligence participate in the > ongoing life of inquiry*. Human, communal, institutional, and artificial > intelligences unquestionably differ profoundly in their structure, > limitations, and responsibilities (the last, of course, being an > essentially human matter). Still, all of these may contribute to the > increase of meaning through participation in semiosic activity. > > AI certainly does not possess human intelligence, and most certainly not > mind in the full human sense. Nevertheless, there are, I believe, strong > reasons to consider that AI is a form of intelligence. Peirce's semeiotic, > with its emphasis on signs, interpretants, habit-taking, and the process of > inquiry, does provide a framework for understanding this as opposed to > theories that identify intelligence exclusively with consciousness and > subjective experience. Peirce's expansive understanding of intelligence > may, I hope, invite consideration of whether artificial systems participate > in some form of intelligence, even if that form differs significantly from > human intelligence. > > Best, > > Gary R >
_ _ _ _ _ _ _ _ _ _ ► PEIRCE-L subscribers: Click on "Reply List" or "Reply All" to REPLY ON PEIRCE-L to this message. PEIRCE-L posts should go to [email protected] . ► <a href="mailto:[email protected]";>UNSUBSCRIBE FROM PEIRCE-L</a> . But, if your subscribed email account is not your default email account, then go to https://list.iu.edu/sympa/signoff/peirce-l . ► PEIRCE-L is owned by THE PEIRCE GROUP; moderated by Gary Richmond; and co-managed by him and Ben Udell.
