Please see some of my essays; consider them to be drafts.
Visit also my Renaissance Philosophy Blog

Saturday, August 15, 2009

Michael Polanyi and His Debate with Alan M. Turing

The Immortality of the Intellect Revived:
Michael Polanyi and His Debate with Alan M. Turing

By Paul Richard Blum

[Now pp. 269-279 in Das Wagnis, ein Mensch zu sein: Geschichte - Natur - Religion

and in

Knowing and Being: Perspectives on the Philosophyof Michael Polanyi. Edited by Tihamér Margitay. Cambridge Scholars Publishing. 12 Back Chapman Street, Newcastle upon Tyne, NE6 2XX, UK. ISBN (10): 1-4438-2062-8]

For a machine is a machine only for someone who relies on it (actually or hypothetically) for some purpose, that he believes to be attainable by what he considers to be the proper functioning of the machine: it is the instrument of a person who relies on it. (Polanyi 1974:262)

When I read this passage in Michael Polanyi's Personal Knowledge, it struck me as a proof of the immortality of the soul by way of operation. The purpose of my paper is to explain that. I will first explain the context and importance of this passage within Polanyi's philosophy of cognition, and then I will explain the context of the debate about the immortality of the soul.

The quoted passage appears in the chapter "The Logic of Affirmation". The philosopher starts with considerations concerning the usage of language, especially the distinction between the confident or direct use of a word and the skeptical or oblique employment of language. What we know elsewhere as the relationship between language and metalanguage (Polanyi will refer to Tarski in due course) is presented here as a distinction necessary for proving the existence of personality in knowledge as being manifest through assent to the "substantial character" of any utterance. Hence follows his emphasis on the awareness ("watching") that takes place in the act of speaking: "The formalization of the meaning relies therefore from the start on the practice of unformalized meaning." (250) One could think that Polanyi puts a wedge between the mind and its objects (i.e., everything that is intended, understood, and referred to). Of course that's not appropriate: Polanyi circumscribes the process of understanding through language as a transition from the non-formalized towards the form. Now, 'the form' is obviously not some kind of Platonic idea or Aristotelian essence, it must be the process -- or the product of the process -- of the buildup of meaning.[1]

On the side of the thinking subject we have the "act of confidence", i.e., that which makes knowledge personal, as the book again and again explains. Words are used obliquely, when we elevate the object of our utterance to metalanguage -- in writing visible by putting quotation marks around terms. This skeptical use of words is a reflective act intended to repair the unavoidable insecurity of confident utterance. For the confident utterance always runs a risk of misplacement (250-251). Therefore confidence and insecurity coincide in the person who dares to put forward a statement. In the next step, Polanyi describes the strategies of oscillating between direct and oblique speaking as loops ("indefinite and futile regress", 252) that never can escape the basic fact, according to which language appears to be a tool of the act of confidence: "… only a speaker or listener can mean something by a word, and word in itself can mean nothing" (252). Polanyi observes that the loop to capture precision by means of words is "logically meaningless" and a fallacy for the very reason that words are instruments of meaning but neither the meaning itself nor the person that means, whatever is meant (252). At this point I may remind of Nicholas of Cusa (1401-1464) who described with his own terminology the impossibility to attain precision with the same concepts that need to be sharpened.[2]

In the next chapter we learn that in a critical epistemological operation the act of an assertion-of-fact may be tentatively separated from the asserted fact. Polanyi suggests to separate the act from the fact in an assertion and to compare the fact with experience. However, he does not speak of 'fact' but of the sentence uttered in an assertion. Therefore he is not endorsing an empirical verification but invites to exercise a phenomenological process, namely, to watch the assertion as-though-it-were-not-asserted. The content of the affirmation is scrutinized without the act. "If as a result of this test we decide to renew the act of assertion, the two parts are reunited and the sentence is reasserted." (254) Obviously, such verification is a self-referential process rooted in the knowing person: "It is an act of tacit comprehension, which relies altogether on the self-satisfaction of the person who performs it." (254) Assertion, in so far as it produces a statement held to be true, is a personal act that nods towards a state of affairs that is given in a non-stated statement, after assent had been withheld. However, we should scratch out any temporal notion, for phenomenologically speaking the mental experiment described here is the essence of assent and not its gradual product. Conversely, we may infer, negation is the sustained withholding of the tacit comprehension.

After these premises there is no doubt that thinking, affirming, or claiming anything, originates in a fuzzy subjectivity, called the person, that shapes itself in the act of understanding, uttering, affirming. At this point Polanyi addresses the question of thinking machines. This question was a highly debated issue in the middle of the 20th century. In his terminology, as said before, formalization is the act of shaping that, which is understood, and presupposes a level that is itself unformalized. He now describes the operation of a computer as follows:

… a formal system of symbols and operations can be said to function as a deductive system only by virtue of unformalized supplements, to which the operator of the known system accedes: symbols must be identifiable and their meaning known, axioms must be understood to assert something, proofs must be acknowledged to demonstrate something, and this identifying, knowing, understanding, and acknowledging, are unformalized operations on which the working of the formal system depends.... These are performed by a person with the aid of the formal system, when the person relies on its use. (258)

In this train of thought, a computer is de facto defined as a formalized version of operations, which in an originary way are prior to formalization. "The legitimate purpose of formalization lies in the reduction of the tacit coefficient to more limited and obvious informal operations." (259) As Polanyi explains elsewhere in the book, 'the tacit coefficient' is the contribution of the person in the act of knowing, meaning, etc., for instance, the confidence in a statement. Therefore, the act of entrusting calculations to a machine is nothing but "shifting confidence" (260) from the personal utterance to the correctness of the operation as formalized thinking. It is at this point that Polanyi made the statement quoted at the beginning: "For a machine is a machine only for someone who relies on it (actually or hypothetically) for some purpose, that he believes to be attainable by what he considers to be the proper functioning of the machine: it is the instrument of a person who relies on it." He further explains it by the following set of parallels[3]:






functions, purposes etc. entertained by the mind


Neurological model of subject

Intellectual purposes attributed to the subject by the neurologist

In this analogy between brain and machine it is visible that computing, i.e., the calculating operations of the computer, are not only presented as thoughts entertained by a mind, but as the operations of the brain. 'Mind' in the computational sense can be seen as mechanical functions only if investigated by a thinking subject that is not the thinking mind itself: for instance a neurologist. Neuroscience looks at the mind as though it were a machine, and not the other way around, at a machine that operates itself as though it were a mind. To treat a computer as though it were a mind was, indeed, the challenge of the notion of 'thinking machines'. We have not yet arrived at immortality, but we have reached the interface between machine and mind that was important to Polanyi and his contemporaries. His statement underscores the dependence of the machine on the subject, which itself may be a machine, if and only if looked at it in this mode and perspective. As we saw, his motivation is logical (and of course also anthropological) in that he emphasized (through the mental experiment of withholding affirmation) the indeterminacy of the mind in judging and exercising its operations. Polanyi comments upon the tripartite diagram with the explicit warning that the neurologist's focus of interest does not at all represent personal "functions of the subject's mind" (262-263). The mind operates in the act of cognition in the same way as words acquire meaning in the act of speaking, but the analogy points to the unspecified level both in thinking and in speaking before the understanding or before meaning.

On October 27, 1949, an interdisciplinary discussion on "The Mind and the Computing Machine" took place at Manchester University in England (Hodges 1983: 415). A typescript of only five pages summarized some of the issues.[4] One of the participants was Alan Turing, who maintained that a "machine may be bed with incompatibles, but when it gets 'contradiction' as a result, there is then a mechanism to go back and look at things which led to the contradiction." We may note that he is describing the operation to withhold an assertion and compare the outcome as though it were not asserted, and -- in this case -- to deny the sentence, which then is the same as to decide not to repeat the assertion. A participant of this discussion, Geoffrey Jefferson, objected that this is (a) an "argument against the machine", probably because he thought that a machine is not able to go back of itself to the origin of an incompatibility. He further doubted that even intelligent beings would necessarily perform this control act; he asked (b) whether "human beings do this kind of thing?"[5] To that Turing answered: "yes -- mathematicians". Turing is on the way to mechanize human thought: he concedes that only mathematicians would care to repeat the operation backwards in order to find out the flaw in the calculation after having maintained that a computer would do what Descartes had strongly suggested to do.

It is a Cartesian rule to view inferences as operational chains. However, we know what followed – rationalism wed with mechanicism. To Descartes the mind was a ghost in a machine. In Polanyi's terms, Turing answered like a neurologist who switches his attention from thinking to objectifying mind and its outcome.

The transcript of the discussion reports that after Turing had confirmed that, indeed, mathematicians do check the operations for flaws, someone muttered: "are mathematicians human beings?" This was probably meant as a joke but it mimics the reversal of perspective, for Turing now explained that he was thinking of "the kind of machine which takes problems as objectives, and the rules by which it deals with the problems are different from the objective," thus clearly aiming as dissecting the problem into manageable units: problem-objective-rule. He explicitly referred to Polanyi who had made the distinction "between mechanically following rules about which you know nothing, and rules about which you know." At this point Polanyi intervened by saying that the rules of conduct or behavior are different from logical rules. Turing now emphasized the parallel between conscious working and routine: "a machine may act according to two different sets of rules", the conscious method vs. the habitual method. Although there seems to be consensus so far, Polanyi now doubted that "the semantic function can ultimately be specified"; and he stated firmly: "whereas in point of fact a machine is fully specifiable, while a mind is not." The so-called specification must mean the illusion that the computing entity (brain or machine) can be fully programmed and therefore, by extension, also the mind. Turing replied "that the mind is only said to be unspecifiable because it has not yet been specified; but it is a fact that it would be impossible to find the programs inserted into quite a simple machine -- and we are in the same position as regards the brain. The conclusion that the mind is unspecifiable does not follow." (Emphasis in the transcript.) He seems to make a very important concession, namely, that it is impossible to find the program a posteriori, while he believes it to be possible to program the machine completely, a priori. Polanyi answered that this "should mean that you cannot decide logical problems by empirical methods. The terms by which we specify the operations of the mind are such that they cannot be said to have specified the mind. The specification of the mind implies the presence of unspecified and pro-tanto unspecifiable elements." He finally adds that "acceptance as a person implies the acceptance of unspecified functions".

It is well known that this dialogue prompted Turing to publish his famous paper on "Computing Machinery and Intelligence" (1950; Turing 1982) suggesting the later so called Turing Test, as it is obvious that it guided Polanyi's treatment of the question of acceptance in his book on Personal Knowledge. There the author refers on purpose to Turing's "speculations" that made him famous. Polanyi translates Turing's opening question "Can machines think?" into "the experimental question, whether computing machines could be constructed to deceive us..." (Polanyi 1974:263 note 1). Without entering the debate about the Turing Test it is sufficient to say that Turing makes a machine

· suspend its judgment, when it returns to previous stages of an operation if the result of it turns out to be contradictory

· operate upon itself, if it is fully 'specifiable'

· transform a posteriori experience into purposes.

All those pertain to 'personal knowledge' in Polanyi's view.

In order to come to the immortality question, it must be emphasized that Turing and Polanyi agree upon a certain dualism, in Turing's terms the distinction between conscious and automatic. It is clear that automatic actions are such that they work apparently the same way as conscious actions -- only without thinking.[6] Remember that Aristotle had stated as a fact of definition that technique does not deliberate and that a musician comes to perfection through practice, rather than deliberation.[7] Remember also Thomas Aquinas's proof for the existence of God, The Fifth Way, where he says that an arrow hits the target as though it were purposefully going after it (Summa th. I 2.3). Automatic action without consciousness is nothing new in philosophy. Where Turing and Polanyi disagree is the possibility -- even in theory -- to program a machine so that the machine operates responsibly in a variety of senses of the word.[8] Polanyi emphasizes that behind every determination is a realm of indeterminacy, behind every form is what is forming. It is hard to avoid at this point to see 'the person' that Polanyi invokes in terms of the absolute.[9] Given the possibilities and options of mechanized thinking the question at stake is whether thinking is something mechanical.

Modern thought since the revival of the geometric method is not anymore described as a correspondence between things and thoughts, but as operations that are controlled -- by whomever. Although the question of accordance between propositions and states of affairs has been debated again and again, the means to control and assess such concord are always operational. This is the case with Descartes's four rules of inference, this is the case with modern geometry that does not inquire into the essence of geometrical proportions but into their construction, as Kant among others has convincingly shown. The fallacy committed by Turing, at least from the point of view of Polanyi, consisted in transferring the orderliness of the operation into the essence of the operation. Going back to the origin of the 'contradiction!' is the operational aspect of assent or denial. Polanyi stresses that the operation is not its own origin. This is where the dualism has to be found. Traditionally speaking this dualism is that between mind and matter, which dominated the debate about the immortality of the soul.s

The question of the immortality of the soul has an ontological and an epistemological side. Of course, it had always been interesting for the ontological aspects, the question being, whether humans have an eternal life. However, epistemology has been the battlefield of this debate, because it is the operation of the mind/soul that at best manifests features that allow conjectures of immortality (cf. Blum 2007 with copious reference to primary and secondary sources). The question, as everybody knows, goes back to Aristotle, who had pondered the option that the human intellect could be separable from the rest of the soul and the living being (De anima III 5), thus bestowing the thinking power with the ontological status of a spirit in materially untainted action. The mixture between metaphysics and epistemology continued into the Renaissance, with added urgency from Christian doctrine. Those were the most common solutions to the problem:

Platonism: the human mind is the interface between the spiritual and the corporeal worlds, but it stays on the spiritual side. The corporeal world may be what it is, at any rate, the mind cognizes bodily things by way of recognition of forms in the mind itself that correspond in some way to the objects seen (Ficino 2001-2006, vol. 2; Platonic Theology lib. 8). This is in brief terms innatism. The epistemological problem is obviously the impossibility to verify the objects that are supposed to be cognized. However, the soul is exclusively busy with perfecting itself and purifying itself from any contingency (and it is only purification why the mind engages in cognition). Therefore the critical scientific question cannot be addressed with Platonism. But immortality is guaranteed, because the mind remains un-entangled with mortal objects, and especially if it purifies itself it works its way up to the level of Angels and other immortal beings. The Platonic answer to the question of immortality, therefore, is an essentialist answer, whereby the description of the functions of the mind appears to be purely apologetic. Still, there is some epistemological merit, if the cognitive operations are analyzed as shaping sensual reality.

Aristotelianism: the human mind processes data delivered through the senses. The question is: Can the human mind think anything that has not been delivered by the lower functions of the soul? Two answers are possible: no, the mind processes only sense data, and thinking is nothing but building up higher and higher levels of complexity in ideas that allow understanding reality in more and more complex ways, to the effect that the illusion of an independent life of the mind can arise, but still is an illusion. This is in bold strokes John Locke's sensualism. In terms of essence it is clear that the human mind is contiguous with the soul and the soul is contiguous with the body, basically a surface phenomenon of bodily functions. The soul is not the root of things but the most complex matter possible. In terms of epistemology, the argument is that if the mind were not cognate with the senses, and if the senses were not bodily, then nothing could be known. This is the solution entertained by Aristotelians like Pomponazzi (1525), although he was not alone. The second answer to Aristotelianism is yes, the mind thinks not only sense data but also purely intellectual objects. To be sure, even the sense data are not processed as such, but as abstractions. Therefore sense perception does not provide a valid argument against the purely spiritual nature of mind.[10] This answer is in reality a hidden Platonism. What needs to be clarified is: where does the mind get its criteria that allow processing all those wonderful sense data? Skepticism and criticism ensued.

The basic distinction between the solutions of the immortality problem is that between an essentialist versus an operationalist approach. The Platonists defined the soul a priori as spiritual and therefore as immortal. All the pains, taken by Marsilio Ficino in 18 books of the Platonic Theology to prove the immortality of the soul, were invested in proving that the soul is spiritual. However, many of his arguments circumscribed the operations of the soul with heightened attention to its cleanness of any material impact. Therefore, it is still safe to say that the Platonists approach is essentialist: whatever the mind does, it is spiritual by definition, and hence immortal.

The operationalist interpretation of the soul entangled the soul hopelessly in matter. It was the way of operating of the mind that seemed to prove that the mind is not separate from its body and hence subject to its fate. As I said, the compromise solution, although given already in Aristotle, according to which the mind is well able to process immaterial thoughts, does not gather its argument from the operation of the mind, but from it substance.

One more complication came into the equation, namely the distinction between actuality and potentiality in Aristotle's psychology. In the famous passage of On the Soul III 5, he had speculated that, if the intellect is in potentiality to its objects, there might be an actual intellect. Arabic interpreters labored much about this passage. For our purposes we need not enter the debate of the unity of the intellect, namely the question whether there is one and only one potential/actual intellect, of which all individual minds are nothing but instantiations.[11] Many philosophers distinguished within the human intellect an active and a passive or potential part. The passive part is the one that receives and processes information from the lower parts of the soul. It does so, however, under the guidance and control of the active intellect (cf. Blum 2007: 221). I hope it is obvious what I'm aiming at. The mechanicist interpretation of mind, which is the one that lends itself to computerization, focuses exclusively on the passive intellect of medieval branding. Polanyi's objection aims at vindicating the active intellect. The active intellect is active in so far as it controls the operation of the passive intellect, and since the passive intellect has to be tainted by the objects it processes, namely, sense data, the active intellect sets the rules of the game without being involved – in Polanyi's terms, without being specified.

Now, the whole pre-modern and early modern debate about the active and passive intellect was aimed at securing immortality to the individual. As the Platonists had established, mind can only be immortal if substantially independent of senses. The Aristotelian theory of the agent intellect was always vulnerable to the operationalist attack, which identified the intellect with the machine that processes information. The only way out is to show that in thinking about things there is a component that not only remains independent in some way, but actively performs the operations that look as though they were mechanical. That's what Polanyi did.

Under the conditions of Kantian criticism, once it was liberated of its radical skeptic and of its psychologist garb given to it by Neo-Kantianism, and under the conditions of phenomenology that scrutinizes the availability of reality in thinking, it was possible for Polanyi to aim at the core of the mind without recourse to psychologism and without lapsing into the mechanistic fallacy that had haunted Aristotelianism. For Neoplatonists the mind was immortal in spite of its operations. For Aristotelians the operations of the mind make it look like matter. Turing seems not to have been aware of the problem, for he seems to think of consciousness according to the model of a loop, a control procedure that needs to be programmed. In the same way as the Platonists would have told the Aristotelians that their epistemology misses the point in the metaphysics of the spirit, in the same way Polanyi appreciates the computerization of mental processes but warns against overstretching its claims. It is exactly because of the operation of the mind, which follows a controllable set of rules, that Polanyi emphasizes personality. Personality means: a priori independence from the rules set and authorship of the same rules are necessary for explaining the operation of the human mind. In being independent it must be free, which traditionally is termed to be immortal.[12]

With these considerations I hope to have shown that the debate between Turing and his opponents repeats a problem that was important in the history of the philosophy of mind or intellection. Polanyi in responding to Turing's challenge exposes to the reader the mixture of metaphysical and epistemological claims included in the seemingly technical question: "Can computers think?" and he brings the traditional solution (operationalist vs. essentialist) further by showing that operation insofar as it is intellectual originates from beyond the area and objects of operation (brain or computer). Polanyi thus helps interpreting (for those interested) the immortality debate of early modernity as a fallacy that identified the thinking subject with its means of thought, although a brain is a brain only for a mind that thinks with it.

Works cited:

Blum, Paul Richard. 2007. "The Immortality of the Soul", in James Hankins (ed.). The Cambridge companion to Renaissance philosophy. Cambridge, UK: Cambridge University Press, 211-233.

Bruno, Giordano. 1998. Cause, principle, and unity, ed. Richard J. Blackwell. Cambridge, U.K.: Cambridge University Press.

Cajetan, Tommaso de Vio. 1514. Commentaria in libros Aristotelis De anima. Venice: Scotus.

Cusanus, Nicolaus. 2001. Complete Philosophical and Theological Treatises, trans. Jasper Hopkins, Minneapolis: Banning (2001), 2 vols. (

Ficino, Marsilio. 2001-2006. Platonic theology, ed James Hankins. Cambridge, Mass: Harvard University Press, 6 vols.

Polanyi, Michael. 1974. Personal Knowledge. Towards a Post-Critical Philosophy, Chicago: University Press [1958/1962].

Pomponazzi, Pietro. 1525. Tractatus acutissimi, utillimi et mere peripatetici. Venetiis, Scotus, 1525; reprint Casarano: Eurocart 1995 (fol. 41r-51v: De immortalitate animae).

Popper, Karl R. "Language and the Body-Mind Problem". Proceedings of the XIth International Congress of Philosophy. Brussels, August 20-26, 1953, vol. 7, Amsterdam/Louvain: North Holland/Nauvelearts, 1953, 101-107 [reprinted in his Conjectures and Refutations, 1963:293–298].

Turing, A. M. 1936. "On computable numbers, with an application to the Entscheidungsproblem." Proceedings of the London Mathematical Society, Series 2, Vol.42 (November 1936) 230-265 (

Turing, A. M. 1982. "Computing Machinery and Intelligence." in Douglas R. Hofstadter and Daniel Clement Dennett (eds.). The mind's I: fantasies and reflections on self and soul. Toronto: Bantam Books, 53-67.

[1] Polanyi 1974:250: "The formalization of meaning relies therefore from the start on the practice of unformalized meaning."

[2] Cusanus 2001, Idiota de mente 3, p. 542: "By means of a very lofty intellectual grasp, enfold into a coinciding both naming and being named, and all will be clear. For God is the Preciseness of whatsoever thing. Hence, if someone had precise knowledge of one thing: then, necessarily, he would have knowledge of all things. Likewise, if the precise name of one thing were known, then the names of all things would be known, because there is no preciseness except with God. Hence, if anyone attained unto a single instance of precision, he would have attained unto God, who is the Truth of all knowable things." Precision is the overall topic of this work.

[3] Polanyi 1974:262; I combined two tables into one.

[4] I used the version available at -- this site is maintained by Andrew Hodges.

[5] Jefferson as a neurosurgeon seems to have entertained an ambiguous anthropological position, for he had advocated frontal lobotomy (which seems to imply mechanistic approach to brain/mind, as described by Polanyi) but also had famously praised nobility and infinity of humankind in a 1949 speach (Hodges 1983: 405).

[6] Cf. Turing's definition of an automatic machine Turing 1936, p. 232, § 2: "If at each stage the motion of a machine … is completely determined by the configuration, we shall call the machine an “automatic machine” (or a-machine). For some purposes we might use machines (choice machines or c-machines) whose motion is only partially determined by the configuration … . When such a machine reaches one of these ambiguous configurations, it cannot go on until some arbitrary choice has been made by an external operator." Obviously, choice is due to ambiguity and made by a thinking subject that is not the machine itself.

[7] Physics II 8, 199 b 28, and Nicomachean Ethics I 6, 1098 a 9 and II 1, 1103 a 34. Giordano Bruno (Bruno 1998, De la causa II, p. 41) had stressed the absence of deliberation by paraphrasing: "This is what Aristotle himself shows with the examples of the perfect writer or perfect lute player … : for great musicians and writers pay less attention to what they are doing than their less talented colleagues, who, because they reflect more, produce work that is less perfect and, what is worse, not free from error."

[8] Compare also Karl Popper's description of a machine, which perhaps was an indirect reply to Turing: "A wall-thermometer may be said not only to express its internal state, but also to signal, and even to describe. … Yet we do not attribute the responsibility for the description to it; we attribute it to the maker. … We do not argue with a thermometer." (Popper 1953:104-105).

[9] The last two paragraphs of Personal Knowledge seem to invoke a super-individual 'person' evolving over time (perhaps akin to Pierre Teilhard de Chardin) "a cosmic field which called forth" all individuals perhaps not different from the Christian God. ("God" is he last word of the book.)

[10] This position was defended by Cardinal Cajetan (Thomas de Vio) in commenting upon Thomas Aquinas, Summa th. I 75-76 (Cajetan 1514: lib. 3).

[11] Polanyi 1974:405, could be read as a version of this, averroistic, notion: "… the centres of the phylogenetic fields of which individuals are offshoots … But we do know that the phylogenetic centres which formed our own primeval ancestry have now produced … a life of the mind which claims to be guided by universal standards." This is beyond the purpose of the present paper.

[12] Polanyi 1974:405: "… the centres of the phylogenetic fields of which individuals are offshoots … may endure for ever … ."

1 comment:

  1. A very thought-provoking article!

    I like the way you connect Polanyi's view of the personal coefficient of knowing with the agent (active) intellect.

    Is there a typo in the first quotation from Turing? " ... machine may be bed with incompatibles, but when it gets 'contradiction' as a result, there is then a mechanism to go back and look at things which led to the contradiction." I can't understand the word "bed". Bred? Embedded? Bedded? Led? Fed?