Sunday, March 09, 2014

Can non-human agents using computer language exemplify "knowing about the world?"

[....] “But the general question of whether non-human agents using the Wolfram Language, or any other computer language, ‘know about the world’ in some suitable sense, that one still stands.”

My response:

I understand knowledge, with Raymond Tallis, as fundamentally a mode of explicitness, of explicit-making consciousness. To elaborate a bit: after Grice, and in the words of Raymond Tallis, “linguistic meaning in the real world does not reside in the behavior of the symbols or expressions of which languages are composed—they are not located in ‘the system of symbols’ or its component terms—but in people who use languages to mean things, and the worlds they live in. This is because the specification of linguistic meanings requires that they are meant (by someone). What is more, in order that I should be able to determine what you mean, I have to intuit what you mean to mean.” This involves, as Searle shows, getting a listener to recognize my intention to communicate just those things I intended to say in the act of communication. One cannot ignore the speaking subject: “Our utterances are invested with, and exploit, an ‘implicature’ in virtue of which we can always imply more than we say. Verbal meaning, in short, resides in acts performed by human being who draw upon their knowledge of the world and make presuppositions about the knowledge possessed by their interlocutors.”

If one believes, as I do and again with Tallis (among others), and yet again after Grice (or Searle for that matter) that “[m]eaning cannot be separated from the psyche of the one who emits meaning, or from the psyche of the one who receives it,” and that our concept of knowledge is intimately tied to the various forms of memory (e.g., factual, experiential, and objectual), to emotions, thoughts, beliefs, and imagination, “the general question of whether non-human agents using the Wolfram Language, or any other computer language, ‘know about the world’ in some suitable sense” lacks any standing whatsoever. The question makes sense only if one thinks of meaning (which is, as Tallis says, ‘a quintessential feature of human consciousness’) “in purely linguistic terms and language being primarily a system of symbols.” One, it seems, has to have a (or something like a) “computational theory of mind” to imagine a computer language might exemplify having knowledge about the world (the relevant ‘knowledge’ here can only be metaphorical or secondary and derivative, parasitic in meaning on the knowledge possessed by those who program the computers, etc.). In short, knowledge requires “an enworlded self.” More explicitly:

“Knowledge begins with the sense of there being something beyond how things appear to us: it begins with the concept of an object that is other than the self who entertains the notion of an object. Implicit in the idea of the object is the intuition of the subject contrasted with the object; more precisely, the Existential Intuition ‘That I am this….’ [the nature and origin of which are discussed in Tallis’s 2004 volume, I Am: A Philosophical Inquiry into First-Person Being] Object knowledge [even Kleinian ‘internal objects’!] is also permeated [as ‘Wittgensteinians’ remind us] by a sense of publicness—of a shared world—that is not available to asocial sentience or asocial neural activities [or an electronic device that performs high-speed arithmetical and logical operations].”

Intentionality is a feature of perceptions, of propositional attitudes such as beliefs and desires, and of utterances such as assertions. This necessarily implicates consciousness, consciousness of something…. Computers are without minds, the most conspicuous feature of which is consciousness. And consciousness cannot be reduced to material or biological or neurological properties: in other words, materialism cannot account for the “indexicality of human consciousness” in the sense of being “here” and “now” as Tallis says, similar to the Da-sein Heidegger identifies as the essence of the human being (Tallis provides compelling arguments against attempts to neurologize ‘here’ and indexicality in general). Computers by definition can’t have first-person experience: a “narrative center of gravity” requires the higher-order activity of a self…. 

See “The Chinese Room all over again? by Catarina Dutilh Novaes at the New APPS blog. 
Update: Professor Dutilh Novaes has replied to my comment as follows: 
 “But to stipulate that intentionality must be exclusively to humans from the start is to beg the question on precisely what is at stake, i.e. can non-human agents instantiate phenomena that are relevantly similar to human cognition? That's one of the points eloquently made by M. Boden in the paper I linked to above.” 
My response: 
Perhaps I’m obtuse, but I fail to see where Boden “eloquently makes that point.” A computer can only instantiate phenomena that are relevantly similar to human cognition to the extent that it is human beings who program computers, and “similar” is then only used rather loosely if not figuratively: For instance, we sometimes hear it said that computers “follow rules,” but computers
“cannot correctly be described as following rules any more than planets can correctly be described as complying with laws. The orbital motion of the planets is described by the Keplerian laws, but the planets do not comply with the laws. Computers were not built to ‘engage in rule-governed manipulation of symbols,’ they were built to produce results that will coincide with rule-governed, correct manipulation of symbols. For computers can no more follow a rule than a mechanical calculator can. A machine can execute operations that accord with a rule, provided all the causal links built into it function as designed and assuming that the design ensures the regularity in accordance with the chosen rule or rules. But for something to constitute following a rule, the mere production of a regularity in accordance with a rule, is not sufficient. A being can be said to be following a rule only in the context of a complex practice involving actual and potential activities of justifying, noticing mistakes and correcting them by reference to the rule, criticizing deviations from the rule, and if called upon, explaining an action as being in accordance with the rule and teaching others what counts as following a rule. The determination of an act as being  correct, in accordance with the rule, is not a causal determination but a logical one. Otherwise we should have to surrender to what results our computers produce.” (Bennett and Hacker) 
The use of language that suggests, for instance, that computers instantiate phenomena “relevantly similar to human cognition” is fairly harmless until it is taken literally, leading us to suppose that it is a fact, or simply possible, that “computers really think, better and faster than we do, that they truly remember, and, unlike us, never forget, that they interpret [or understand] what we type in, and sometimes misinterpret [or misunderstand] it, taking what we wrote to mean something other than we meant. Then the [computer] engineers’ [or scientists’] otherwise harmless style of speech ceases to be an amusing shorthand and becomes a potentially pernicious conceptual confusion,” as is, I think, the case here.   
Dennett would have us speaking of Deep Blue as “playing” chess, just like Kasparov, but the computer only “’plays’ chess in the sense that the microwave ‘cooks’ soup, though the programming is vastly more complicated.” (Daniel Robinson) What’s “stipulative” is the “intentional stance,” fashioned, in part, so as to make it appear plausible that machines (among other things) are, like us, “intelligent systems.” In Robinson’s words, “[c]onsider the broad, various, cultural, and dispositional factors that need to be recruited in order to qualify an activity as ‘play,’ and then array these against whatever ‘process’ gets Deep Blue to have the Bishop move to QP3.” And then, relatedly and further, we might ask, “If Spassky and Kasparov are doubtful as to whether computers are ‘playing’ chess, is it not Dennett who must rethink the matter?” 
It’s on the order of a category mistake to think intentionality applies to non-human agents (although it applies in some degree to at least some non-human animals), Dennett’s “intentional stance” and nonsense about the fictional character of folk psychology notwithstanding: the ascription of psychological attributes is not about an interpretative stance, heuristic overlays or theoretical posits (it’s not surprising that Boden uncritically cites Dennett on this score). One does not merely adopt an “intentional stance” in the use of psychological predicates.* But my point concerns consciousness (intentionality being one feature or property of consciousness) in the first instance and not intentionality, at least insofar as some mental phenomena are not obviously intentional in any conventional sense (e.g., moods or sensations). In any case, it would be more precise to say, after Bennett and Hacker, that what is intentional is “the psychological attribute that has an intentional object.” Therefore,
“[o]ne cannot intelligibly ascribe ‘intentionality’ to molecules, cells, parts of the brain, thermostats or computers. Not only is it a subclass of psychological attributes that are the appropriate bearers of intentionality and not animals or things, but, further, only animals, and fairly sophisticated animals at that, and not parts of animals, let alone molecules, thermostats or computers, are the subjects of such attributes. …[I]t makes no sense to ascribe belief, fear, hope, suspicion, etc. to molecules, [contra Searle] the brain or its parts, thermostats or computers.” 
* For the full critique of Dennett on this score, see the first appendix to M. R. Bennett and P.M.S. Hacker’s Philosophical Foundations of Neuroscience (2003). I agree with Tallis who writes, “It is difficult to know why this argument has been taken seriously.” See too the debate in Maxwell Bennett, Daniel Dennett, Peter Hacker, and John Searle (with Daniel Robinson), Neuroscience and Philosophy: Brain, Mind, and Language (2007).
Further Reading: 
  • Bennett, M.R. and P.M.S. Hacker. Philosophical Foundations of Neuroscience. Malden, MA: Blackwell, 2003. 
  • Bennett, Maxwell, Daniel Dennett, Peter Hacker, John Searle, and Daniel Robinson. Neuroscience and Philosophy: Brain, Mind and Language. New York: Columbia University Press, 2007. 
  • Descombes, Vincent (Stephen Adam Schwartz, tr.). The Mind’s Provisions: A Critique of Cognitivism. Princeton, NJ: Princeton University Press, 2001. 
  • Gillett, Grant. Subjectivity and Being Somebody: Human Identity and Neuroethics. Exeter, UK: Imprint Academic, 2008. 
  • Gillett, Grant. The Mind and Its Discontents. New York: Oxford University Press, 2009. 
  • Grice, Paul. Studies in the Way of Words. Cambridge, MA: Harvard University Press, 1989. 
  • Hacker, P.M.S. Human Nature: The Categorial Framework. Malden, MA: Blackwell, 2007. 
  • Horst, Steven. Beyond Reduction: Philosophy of Mind and Post-Reductionist Philosophy of Science. Oxford, UK: Oxford University Press, 2007. 
  • Hutto, Daniel D. The Presence of Mind. Amsterdam: John Benjamins, 1999. 
  • Hutto, Daniel D. Beyond Physicalism. Amsterdam: John Benjamins, 2000. 
  • Hutto, Daniel D. Folk Psychological Narratives: The Sociocultural Basis of Understanding. Cambridge, MA: MIT Press, 2008. 
  • Hutto, Daniel D., ed. Narrative and Folk Psychology. Exeter, UK: Imprint Academic, 2009. 
  • Pardo, Michael S. and Dennis Patterson. Minds, Brains, and Law: The Conceptual Foundations of Law and Neuroscience. New York: Oxford University Press, 2013. 
  • Robinson, Daniel N. Consciousness and Mental Life. New York: Columbia University Press, 2008. 
  • Searle, John R. Intentionality: An Essay in the Philosophy of Mind. Cambridge, UK: Cambridge University Press, 1983. 
  • Searle, John R. The Rediscovery of the Mind. Cambridge, MA: MIT Press, 1992.
  • Tallis, Raymond. The Explicit Animal: A Defence of Human Consciousness. New York: St. Martin’s Press, 1999 ed. 
  • Tallis, Raymond. I Am: A Philosophical Inquiry into First-Person Being. Edinburgh: Edinburgh University Press, 2004.
  • Tallis, Raymond. The Knowing Animal: A Philosophical Inquiry into Knowledge and Truth. Edinburgh: Edinburgh University Press, 2005. 
  • Tallis, Raymond. Aping Mankind: Neuromania, Darwinitis and the Misrepresentation of Humanity. Durham, England: Acumen, 2011.

0 Comments:

Post a Comment

<< Home