(Albert Einstein, slightly but relevantly misquoted.[i])
For us it is the circumstances under which he had such an experience that justify him in saying in such a case that he understands, that he knows how to go on….This will become clearer if we interpolate the consideration of another word, namely “reading.”… The use of this word in the ordinary circumstances of our life is of course extremely familiar to us. But the part the word plays in our life, therewith the language-game in which we employ it, would be difficult to describe even in rough outline. (Wittgenstein, 1953: p. 61e, §§155-156)
‘Twas brillig, and the slithy toves
Did gyre and gimble in the wabe;
All mimsy were the borogoves,
And the mome raths outgrabe. (Carroll, 1988)
TABLE OF CONTENTS
3. Necessary and Sufficient Conditions for Legibility and Reading
4. Legibility, Reading, and The Falsity and Impossibility of Strong AI
This essay will be published in three installments; this third and final installment contains section 4.
You can also download and read or share a .pdf of the complete text of this essay HERE.
4. Legibility, Reading, and The Falsity and Impossibility of Strong AI
As we’ve seen, the intentional targets of the act or process of reading are at-least minimally scannable, at-least minimally parse-able, and at-least minimally comprehensible structural objects belonging to some or another language L, that are ineluctably embedded in an egocentrically-centered, orientable, manifestly real, three-dimensional space, thereby necessarily requiring the actual existence and essential embodiment of the reader. As linguistic structural objects, the intentional targets of reading are manifestly real linguistic physical tokens of manifestly real linguistic physical types, which in turn are inherently repeatable objects that are non-platonically and kantianly abstract according to this definition:
X is non-platonically and kantianly abstract if and only if X is not uniquely located and realized in manifestly real spacetime, and X is concrete otherwise. (Hanna, 2015: pp. 269-270)
Now, the rational human cognition of concrete tokens of the linguistic structural objects of reading, whether in perception, memory, or imagination, is what Kant calls sensibility (Sinnlichkeit), which in turn requires a capacity for first-order conscious or self-conscious, essentially non-conceptual, and non-empirical unified formal spatial or temporal representation, or what Kant calls pure intuition or reine Anschauung (Kant, 1998: p. 173, A20/B34-35). Therefore, the act or process of reading is an essentially intuitionistic activity that doesn’t require any sort of platonic objects. The act or process of reading thereby wholly avoids the classical metaphysical/ontological and epistemic problems of platonism, especially including The Benacerraf Dilemma, which says: (i) on the one hand, our standard Tarskian semantics of mathematical truth requires platonically abstract objects that exist outside of spacetime and are causally inert, but (ii) on the other hand, our best theory of human knowledge requires directly sensibly accessible causal objects of perception, so (iii) mathematical truth is humanly unknowable (Benacerraf, 1973). In short, the act or process of reading, by virtue of its intuitionistic nature, is decisively (to coin a nifty neologism) trans-Benacerraf-Dilemma-istic, precisely because it’s metaphysically structuralist, ontologically non-platonistic, although fully accommodating non-platonically and kantianly abstract objects, and epistemically skepticism-resistant, from the get-go (Hanna, 2015: chs. 6-8).
Furthermore, and now finally getting down to answering the leading question of this essay, my theory of legibility and reading predicts that there are legible texts that ordinary rational human minded animals can read, that even the world’s most sophisticated robot cannot read, even when we bracket temporarily the contested issue of the role of consciousness or subjective experience, i.e., sentience, vs. computational zombie-states, i.e., non-consciousness or non-sentience, in acts or processes of reading.
To show this, let’s consider computational reading that’s based on optical character recognition (OCR), and let’s also make the plausible assumption that even the world’s most sophisticated robot will have to employ some or another version of OCR:
There are two basic methods used for OCR: [m]atrix matching and feature extraction. Of the two ways to recognize characters, matrix matching is the simpler and more common.
Matrix Matching compares what the OCR scanner sees as a character with a library of character matrices or templates. When an image matches one of these prescribed matrices of dots within a given level of similarity, the computer labels that image as the corresponding ASCII character.
Feature Extraction is OCR without strict matching to prescribed templates. Also known as Intelligent Character Recognition (ICR), or Topological Feature Analysis, this method varies by how much “computer intelligence” is applied by the manufacturer. The computer looks for general features such as open areas, closed shapes, diagonal lines, line intersections, etc. This method is much more versatile than matrix matching. Matrix matching works best when the OCR encounters a limited repertoire of type styles, with little or no variation within each style. Where the characters are less predictable, [intelligent character recognition, or topological feature analysis,] is superior. (Data ID, 2023)
Now, let’s consider garbled texts: that is, texts that contain misspelled sub-texts, sub-texts with missing characters, sub-texts with obscured characters, sub-texts whose characters are excessively large or excessively small, ungrammatical sub-texts, incomprehensible sub-texts, and above all, texts that contain disoriented sub-texts, that is, sub-texts reversed in a mirror, tipped sideways, or upside down. Necessarily, any digital computer running an OCR program must process information in a step-by-step sequence, and whenever it encounters something that it cannot recognize as a determinate unit of information, whether by matrix matching, feature extraction, aka intelligent character recognition, aka topological feature analysis, or whatever, it simply stops processing. Now, there is no general algorithm for determining, from a description of an arbitrary computer program and an arbitrary input, whether this program has either actually finished processing, or will start up again and continue to run forever. In the logical theory of digital computing, this is known as the halting problem, and it’s provably unsolvable (Boolos and Jeffrey, 1989: pp. 28-33, 41-42, 49-50).
But, as ordinary rational human minded animals, we intuitionistically represent texts as complete Gestalt-structures that are embedded in manifestly real, egocentrically-centered, orientable space, and therefore we always have a unified formal spatial representation of the text as a whole for guiding us through our reading, not only before we begin scanning it sequentially, but also throughout the time we’re scanning it sequentially. This enables us to jump over, fill in, or creatively interpret illegible sub-texts, and/or re-orient disoriented sub-texts in spatial imagination, when we encounter garbled texts, hence we’re able to read all sorts of garbled texts, provided that they’re otherwise at-least minimally legible by the criteria I provided above. Hence our ordinary rational human minded animal ability to read garbled texts, provided that they’re otherwise at-least minimally legible, will necessarily exceed the digital processing abilities of any and all computers to read those texts, i.e., there are some legible texts that ordinary rational human animals can read, that even the world’s most sophisticated robot can’t read.
Here’s an example of a legible text that any ordinary rational human animal can read, but even the world’s most sophisticated robot can’t read, using a text that we’ve seen twice already:
‘Twas brillig, and the slithy toves
Did gyre and gimble in the wabe;
All mimsy were the borogoves,
And the mome raths outgrabe. (Carroll, 1988)
By hypothesis, this text from Jabberwocky satisfies the perceptibility condition and the syntactic condition, yet also fails the semantic condition. So it’s prima facie illegible and unreadable. Now, consider any ordinary rational human minded animal, and for convenience let’s call that person “Bob.” And correspondingly, let’s consider the world’s most sophisticated robot, a behavioral counterpart to Bob, and for convenience call it “Robobob.” After successfully scanning and parsing that text from Jabberwocky, Robobob attempts to comprehend it, but cannot do so, and concludes that it’s incomprehensible, so stops processing. But Bob, who like any other ordinary rational human minded animal, has an innate capacity for creative “gumption”—i.e., creative initiative and resourcefulness, which of course both Lewis Carroll and Einstein possessed to an extraordinary degree—doesn’t give up, and continues to think about the text, muse about it, sleep on it, and dream about it. Then finally, when Bob wakes up the next day, he finds that, like a creative artist or creative scientist, he’s freely and spontaneously assigned private meanings to all the nonsense terms, and has a novel semantic Gestalt of the entire text, so that the text is now fully legible for him. These meanings are not necessarily private—a “private language” in that absolute sense, as Wittgenstein compellingly argued, is conceptually impossible (Wittgenstein, 1953: pp. 88e-104e, §§242-315; Hanna, 2021, ch. XIII)—since in principle Bob could tell other people about them, or others could somehow learn about these meanings in some other way: hence they’re only contingently private and in-principle universally shareable. But, as a matter of fact, Bob never tells anyone about them, and no else ever learns about them, including of course Robobob. Yet the Jabberwocky text is legible for Bob in all three senses, and he privately enjoys reading it over and over, for the rest of his life. Nevertheless, Robobob cannot read that text because it stopped processing, and also, above all, because it’s nothing but a mobile digital computer and therefore lacks any inherent capacity whatsoever for creative gumption, although of course it could be programmed to exhibit behavior that mimics creative gumption. Hence there’s at least one legible text, i.e., that Jabberwocky text, that’s legible for Bob, and also for any other ordinary rational human minded animal with at least as much creative gumption as Bob, that even the world’s most sophisticated robot can’t read. So the strong AI thesis is false and the strong AI program is impossible.
NOTE
[i] The original quotation is: “The only thing you absolutely have to know, is how to locate the library.”
REFERENCES
(Benacerraf, 1973). Benacerraf, P. “Mathematical Truth.” Journal of Philosophy 70: 661-680.
(Block, 1980). Block, N. (ed.), Readings in the Philosophy of Psychology. 2 vols., Cambridge, MA: Harvard Univ. Press. Vol. 1.
(Boolos and Jeffrey, 1989). Boolos, G. and Jeffrey, R. Computability and Logic.3rd edn., Cambridge: Cambridge Univ. Press.
(Brian, 1996). Brian, D. “Einstein: A Life, Ch 1.” The Washington Post. Available online at URL = <https://www.washingtonpost.com/wp-srv/style/longterm/books/chap1/einstein.htm>.
(Carroll, 1988). Carroll, L. Through the Looking-Glass. New York: Dial.
(Chomsky, 1957). Chomsky, N. Syntactic Structures. The Hague: Mouton.
(Chomsky, 1988). Chomsky, N. Language and Problems of Knowledge. Cambridge MA: MIT Press.
(Data ID, 2023). Data Identification Online. “What’s OCR?” Available online at URL = <http://www.dataid.com/aboutocr.htm>.
(Gödel, 1931/1967). Gödel, K. “On Formally Undecidable Propositions of Principia Mathematica and Related Systems.” In J. van Heijenoort, (ed.), From Frege to Gödel. Cambridge MA: Harvard Univ. Press. Pp. 596-617.
(Hanna, 2006). Hanna, R. Rationality and Logic. Cambridge: MIT Press. Available online in preview at URL = <https://www.academia.edu/21202624/Rationality_and_Logic>.
(Hanna, 2015). Hanna, R. Cognition, Content, and the A Priori: A Study in the Philosophy of Mind and Knowledge. THE RATIONAL HUMAN CONDITION, Vol. 5. Oxford: Oxford Univ. Press. Also available online in preview HERE.
(Hanna, 2018a). Hanna, R. Preface and General Introduction, Supplementary Essays, and General Bibliography. THE RATIONAL HUMAN CONDITION, Vol. 1. New York: Nova Science. Available online in preview HERE.
(Hanna, 2018b). Hanna, R. Deep Freedom and Real Persons: A Study in Metaphysics. THE RATIONAL HUMAN CONDITION, Vol. 2. New York: Nova Science. Available online in preview HERE.
(Hanna, 2018c). Hanna, R. Kantian Ethics and Human Existence: A Study in Moral Philosophy. THE RATIONAL HUMAN CONDITION, Vol. 3. New York: Nova Science. Available online in preview HERE.
(Hanna, 2020). Hanna, R. “Will-Power: Essentially Embodied Agentive Phenomenology, By Way of O’Shaughnessy.” In C. Erhard and T. Keiling (eds.), Routledge Handbook: The Phenomenology of Agency. London: Routledge. Pp. 312-333. Available online in preview HERE.
(Hanna, 2021). Hanna, R., The Fate of Analysis: Analytic Philosophy From Frege to The Ash-Heap of History. New York: Mad Duck Coalition. Affordably available in hardcover, softcover, and Epub at URL = <https://themadduckcoalition.org/product/the-fate-of-analysis/>.
(Hanna and Maiese, 2009). Hanna, R. and Maiese, M., Embodied Minds in Action. Oxford: Oxford Univ. Press. Available online in preview at HERE.
(Hanna and Paans, 2021). Hanna, R. and Paans, O. “Thought-Shapers.” Cosmos & History 17, 1: 1-72. Available online at URL = <http://cosmosandhistory.org/index.php/journal/article/view/923>.
(Hawkins and Allen, 1991). Hawkins, J.M. and Allen, R. (eds.), The Oxford Encyclopedic English Dictionary. Oxford: Clarendon/Oxford Univ. Press.
(Kant, 1998). Kant, I. Critique of Pure Reason. Trans. P. Guyer and A. Wood. Cambridge: Cambridge Univ. Press. (1781/1787, Ak 3, 4: 1-252).
(Kim, 2011). Kim, J. Philosophy of Mind. 3rd edn., Boulder, CO: Westview.
(Putnam, 1983). Putnam, H. “There is At Least One A Priori Truth.” In H. Putnam, Realism and Reason: Philosophical Papers, Vol. 3 Cambridge: Cambridge Univ. Press. Pp. 98-114.
(Russell, 1940). Russell, B. An Inquiry into Meaning and Truth. London: George Allen and Unwin.
(Smith-Garcia, 2020). Smith-Garcia, D. “Einstein Syndrome: Characteristics, Diagnosis, and Treatment.” Healthline. Available online at URL = <https://www.healthline.com/health/einstein-syndrome>.
(Sowell, 1997). Sowell, T. The Einstein Syndrome: Bright Children Who Talk Late. New York: Basic Books.
(Tarski, 1943). Tarski, A. “The Semantic Conception of Truth and the Foundations of Semantics.” Philosophy and Phenomenological Research 4: 342–360.
(Tarski, 1956). Tarski, A. “The Concept of Truth in Formalized Languages.” In A. Tarski, Logic, Semantics, and Metamathematics. Oxford: Oxford Univ. Press, 1956. Pp. 152–278.
(Wittgenstein, 1953). Wittgenstein, L. Philosophical Investigations. Trans. G.E.M. Anscombe. New York: Macmillan.
Against Professional Philosophy is a sub-project of the online mega-project Philosophy Without Borders, which is home-based on Patreon here.
Please consider becoming a patron!
