User:Leucippus/Sandbox2
'Intelligence' explicated Or, Traces To Nowhere[edit]
Caveat[edit]
What lies below is... but a sketch —"hang it in the Louvre?" I hear you cry. No! —I think not. but it'll do for now, as just that, a sketch...I must allow it time to mature, as is natural, before it is granted essayhood... lets hope it matures into a fine cheese. At the moment, this sandbox is spiralling off in various directions but I shall endeavour to bind all these threads together, into a coherent web. With this in mind, this sandbox might alternatively be called "Traces to Nowhere" (in homage to Twin Peaks) that is, until I start the process of binding the stray threads together.
Conventions:[edit]
- 'para' = paraphrase i.e., not a verbatim quote from the source, but only deviating in minor respects
- 'qte' = quoting the source verbatim
- '*' = my notes
Introduction[edit]
Explication[edit]
The ordered pair is a device for treating objects two at a time as if we were treating objects of some sort one at a time, example: As Moe Green put it “Fredo was banging cocktail waitresses two at a time... players couldn’t get a drink at the table.” The ordered pair, in the case, would’ve endowed Fredo with the ability to treat two cocktail waitresses one at a time, thus making his dealings more expedient and, crucially, players would be able to get a drink at the table!
The ordered pair enables us to assimilate relations to classes e.g. the father relation becomes the class of just those ordered pairs which have a male as the first component and his respective offspring as the second component e.g. 〈Abraham, Issac〉 The former component of an ordered pair is interpreted as ‘first’, the latter as ‘second’.
The various definitions of the ordered pair construction e.g:
- 〈x, y〉 ≔ {{x}, {x,y}} or,
- 〈x, y〉 ≔ {{x}, {Ø,y}}
spotlight, in an exemplary way, the conditions for a proper explication.
- The ‘ordered pair’ construction is paradigmatic of what we are up to when we offer an explication of an inadequately formed idea or expression.[note 1]
- Nota bene, explication does not yield synonymy between an explicandum and its explicatum, if it did it would fall prey to the paradox of analysis.
- Explication does not expose hidden meanings (as the words ‘analysis’ and ‘explication’ suggest); we supply lacks.
- It is our job to thoroughly scrutinise this “drive” for hidden meanings. We will find that the apparently indissoluble problems this “drive” produces, can be readily dispersed.
- We focus on the particular functions of the explicandum that make it worth troubling about, and then devise a substitute—an explicatum, clear and couched in terms to our liking, that fills those functions beyond those conditions of partial agreement[note 2] dictated by our purposes and interests, any traits of the explicans come under the head of ‘don’t cares’.
- Explication helps us circumvene the problematic features of ordinary language that lead to puzzling questions and paradoxes. These problems are dissolved in the sense of being shown to be purely verbal, and purely verbal in the important sense of arising from usages that can be avoided in favour of ones that evince no such problems.
- The ordered pair is a paradigm of explication because the function that any explication of ‘ordered pair’ must satisfy can easily be made precise viz. the condition:〈x,y〉=〈w,z〉 iff x = w and y = z. All set theoretic definitions of ordered pair satisfy this condition.
Explication of ‘intelligence’[edit]
- Explication makes a makes a vague concept more explicit.
- I aim to do is provide an explication of the explicandum 'intelligence'. The notion of 'explication', however, has had a long and checkered history in science and philosophy. In general, there are three positions on explication, which make up a bipolar spectrum (note the representatives of each position are not exhaustive): (1) naturalism — represented by Mark Wilson[note 3], W.V. Quine, Otto Neurath, Ernst Mach, and Hume; (2) Carnapian explication (which represents a middle way) - represented by Rudolf Carnap, Hermann von Helmholtz, Kant, and Max Planck; and at the opposite pole (3) constructivism - represented by Thomas Kuhn, Richard Rorty, and Imre Lakatos, Robert Brandom. Now, I will not be discussing positions 2&3 here, but I will stress that I believe the "naturalist" conception of explication - to be the correct one.
- I aim to offer a Quinean or naturalistic explication of 'intelligence'. My explication should be viewed in the larger context of Quine's naturalism (I shall endeavour to make this "context" clear to the reader). Such an explication would involve the regimentation of a concept viz., into first-order logic (hereafter FOL).
- The vague and imprecise explicandum of 'intelligence', once explicated, becomes an explicatum - a more precise, exact, and clear concept. However, there is no sharp boundary between the explicandum of 'intelligence' and its successive explicata - they are continuous with oneanother.
- A naturalistic explication extends the range of application of the original concept. It does this by running together all the explicata into a single concept.
- As a naturalist I believe, pace the constructivists, that concepts are imposed on us by nature and our constructions (such as explications), insofar as they are in accordance with nature, must ultimately be equivalent i.e. they must ultimately not be constructions.
- Depending on temperament, a naturalistic explication can be construed as either "elimination" (qua eliminative physicalism) or, more politely, as "identification" (qua anomalous monism, a form of non-reductive physicalism). To be sure, the vague explicandum of 'intelligence' may have been "eliminated" from scientific discourse, but it still retains its utility for the everyday commonsense-world; the explicatum is to the explicandum, what the microtome is to the chisel. That the vagaries of ordinary language can never be completely eliminated, that is, artificial languages are parasitic upon natural language, albeit in a mutualist way.
- Explication should be viewed as a gradual step-by-step adjustment of human conceptual systems, to a fixed external reality.
My chosen explicandum for 'intelligence'[edit]
- The best general definition of the explicandum 'intelligence', that is, the one that suits the level of generality I am interested in, is the following Wikipedia example:= "Intelligence is the ability or aptitude, to perceive or infer information, and to retain it as knowledge to be applied towards adaptive behaviours within an environment or context". I like this example as it corresponds well to Spearman's "g - factor", which is, perhaps, the best current theory of intelligence i.e. according to a review in the journal Nature:
"(g) accounts for a substantial large amount of variance, it is the source of most of the predictive power of cognitive tests and, as discussed below, it is the locus of most of the genetic variance". — The neuroscience of human intelligence (2010)[1][note 4] |
- NB: My explication of 'intelligence' will, for reasons of brevity and effort, be constrained to the definition provided in this section.
Background work[edit]
Naturalism and Cognitive Science[edit]
“”The evil of the idea 'idea' is that its use, like the appeal in Molière to a virtus dormitiva, engenders an illusion of having explained something.
|
—"The Problem of Meaning in Linguistics", W.V. Quine[2] |
“” The danger of using such undefined and intuitive criteria as pattern, symbol, and logical a prioris, is that linguistics is precisely the one empirical field which may enable us to derive definitions of these intuitive fundamental relationships out of correlations of observable phenomena.
|
—Review of Gray’s Foundations of Language, Zellig Harris[3] |
Source 1: "Whence and Whither the Debate Between Quine and Chomsky?" by Alexander George[4][edit]
qte — "Whence and Whither the Debate Between Quine and Chomsky?"[5]: It is a mistake, however easy, to identify Quine's naturalism with a basic discrimination against mental entities just in virtue of their mentalness. Though pointing to the mental complexion of such entities may be "objection enough", "the naturalists primary objection...is not to meanings on account of their being mental entities"[6];[*For instance, in Quine's essay "Reply to Chomsky" —Quine explicitly states "I may do well to add here an explicit word of welcome toward any innate mechanisms of language aptitude, however elaborate, that Chomsky can make intelligible and plausible"(p. 279), and "Innate mechanism, after all, is the heart and sinew of behaviour"(loc. cit.). On the contrary, see Putnam for sceptical remarks on the kinds of innate mechanisms Chomsky favours — on how hypothetical innate mechanisms can prove wanting in intelligibility when specified in less scrupulously experimental terms than was the concept of quality space.][7]The central complaint (which will be articulated shortly), would remain even if mental terms were somehow construed as having physical denotata.
qte (489): Many of Quine's critics are puzzled by his distinction between underdetermination and indeterminacy. Some have charged that Quine merely makes the uncontroversial claim that current linguistic theory is underdetermined [*e.g. Ricketts, Stabler, Bechtel], that it does not follow ineluctably from a corpus of data, however vast or varied [loc. cit.]. For example P.W. Bechtel asks, "Why, then, should the availability of alternative translation manuals count against our taking a realistic attitude toward one translation manual? As we do with physical theories, why [*can't] we adopt one of these theories despite the underdetermination?"[8] and argues that we can, since, according to him, Quine fails to articulate any deep difference between the two notions.
para (489): Yet two points should be granted to Quine: (1) that an indeterminate theory is not merely an underdetermined one, and (2) that an indeterminate theory cannot be viewed as issuing in statements true or false of the world. The crucial problem is this: An indeterminate theory is underdetermined, but its underdetermination remains (*my emphasis) even after some underdetermined theory has been chosen from each branch of science (current or to come). Where there is slack between observation and theory we have underdetermination, but slack between total theory (viz. all facts, known and unknown) and theory we have indeterminacy. Qte: "If any choice among the many present or future, explanatorily adequate, underdetermined theories of the world would leave unsettled the truth or falsity of linguistics' claims, then we cannot make sense of there being objectively correct evaluations of these" (489). Linguistics was, and still is —alien... to natural science, to our total theory of the world. Ultimately, it is the job, only, of natural science to tell us what is really real; thus, we ought to choose a science independent of intention.
para: Quine declares, [*in the spirit of anomalous monism], that "various complex states of the nerves whereof the neural detail may still be a matter of conjecture... [are the] states that the mental term may be seen as denoting."(sic).[9] Chomsky could quite readily assent to this — since he holds that linguistics — "should sooner or later disappear as a discipline as new kinds of data become available, remaining distinct only in that its concern is a particular faculty of the mind, ultimately the brain"(sic).[10] However, Chomsky believes that, it should be borne in mind that new neurological data (for example) become available in part through a better understanding of the instantiated abstract structures.[11] Such data and accounts of them may be as yet undreamt of in our science, but, as linguistics withers away, we shall gain greater understanding of those neurological facts in virtue of which linguistics' claims hold. In this sense Quine and Chomsky agree about the "level of deepest explanation." Although Chomsky agrees with Quine on the nature of the "deepest explanation," he expects "a serious investigation of behaviour" (RL 24[12]) to yield theories with a distinctive explanatory structure. According to Chomsky,
"...what the organism does depends in part on its experience, but it seems to me entirely hopeless to investigate directly the relation between experience and action. Rather, if we are interested in the problem of 'causation of behaviour' as a problem of science, we should at least analyse the relation of experience to behaviour into two parts: first, LT (learning theory), which relates experience to cognitive state, and second a mechanismcs, which relates stimulus conditions to behaviour given cognitive states CS." (sic) — (RL 26) |
Chomsky's use of nonbehavioural terminology here—(i.e., the postulation of an internal "mechanismcs," and the postulation of a cognitive state "CS")—does not indicate a "difference over what counts as data,"(as Thomas Ricketts suggests)[13]. Rather, it belongs to an advanced stage of inquiry (viz., "learning theory") aiming to solve "the problem of 'causation of behaviour'"—just Quine's problem. On this approach,
"Human action can be understood only on the assumption that first-order capacities and families of dispositions to behave involve the use of cognitive structures that express systems of (unconscious) knowledge, belief, expectation, evaluation, judgment, and the like." (sic) — (RL 24) |
.
In Chomsky's "learning theory," study of a speaker's "cognitive states" (viz., linguistic competence) involves a double departure from a neurophysiological account with behavioural consequences.
para: In the first place, a competence theory will be couched not in neurophysiological terms but in abstract terms that prescind from facts about the neural hardware in virtue of which this abstract description is true of human beings.[note 5]Moreover, grammars of speakers are treated as complex mathematical objects that offer few clues about their physical realization. [note 6]According to Chomsky, that a speaker knows (or has or cognizes) a particular grammar is a fact made true by her physical constitution. Chomsky insists that the structure of the "cognitive state" can be characterised,
"...in principle quite concretely, in terms of physical mechanisms. When I use such terms as 'mind', 'mental representation', 'mental computation', and the like, I am keeping to the level of abstract characterisation of the properties of certain physical mechanisms, as yet almost entirely unknown. There is no further ontological import to mind or mental representations and acts."(sic) — Rules and Representations (New York: Columbia 1980), p. 5. Chomsky, Noam[14] |
Chomsky believes that new neurological data will become available through a better understanding of the instantiated abstract structures.[15]
[*Chomsky believes that all human linguistic behaviour is rule-guided; according to this, even if we have two extensionally equivalent systems of grammatical rules viz., two systems that are true of the same physical object —the brain— they still need not be equally correct: "the right rules are the rules the native speakers themselves have somehow implicitly in mind", and "It is the grammarian's task to find the right rules in this sense."[16][17] Chomsky's position to occupy a mid-point between two conceptions of grammar — "rules as fitting" and "rules as real and overt guides" viz., his decision to construe all grammatical behaviour as "rule guided" by innate structures that are somehow heeded inarticulately; this position is deserving of close methodological attention (attention which Chomsky fails to give).]
para: In the second place, a competence theory requires ascriptions to speakers of states that are only indirectly linked to behaviour. To begin with, not a single datum follows from any competence theory. Without a theory of the (as Chomsky states) "mechanism Mcs, which relates stimulus conditions to behaviour, given the cognitive state CS," nothing about what an individual will say in specific circumstances follows from the ascription of — implicit knowledge of) — a particular grammar to that individual. In particular, it would not follow, as is often assumed, that this speaker will have certain intuitions, that she will make certain judgments about the possible interpretations of presented sentences. In Chomsky's view, a speaker might have terrible intuitions about her language, that is, relatively poor access to the facts about her language which are represented in its grammar.[18] In addition, behaviour will vary not only according to a given human's intelligence, but according to her intentions as well. These are merely particular illustrations of the general fact that knowledge may lead to all sorts of behaviour; no specific behavioural dispositions are intrinsic to possession of any body of knowledge. This "second departure" is independent from the first viz., even if the neurological states in virtue of which someone has a language with grammar G were known, still no dispositions to verbal behaviour would follow from a human's being in those states.
para: Quine might very well tolerate this long-distance relationship between data and theory (*i.e., Chomsky's "double departures" noted previously). After all, Quine understands that the linguist,
"rounds out and rounds off his data. Sometimes the linguist will even reject a heard string as ill-formed, thus rejecting a datum, if he can appreciably simplify his system in so doing; but it would be regrettable to do much of this." — "Methodological Reflections on Current Linguistic Theory"(MRCLT hereafter), W.V. Quine[19] |
Verily, Quine would urge no other course, for this "rounding out and off" is endemic to all rational inquiries.
para: What worries Quine more deeply than the theory's distance from the data is its independence from the data — there is an additional respect in which competence theories stand aloof from behavioural facts. For Chomsky, ascribing knowledge of grammar G to an individual does not merely consist in a recursive specification of those sequences of words which she will find grammatical, nor even in a recursive assignment of <Phonological Form, Logical Form> ordered pairs to such sequences. More readily, it is the hypothesis that that grammar is instantiated neurally in the individual, and not some other grammar [G'] which accords with G in some respects e.g., they both recursively specify the same grammatical strings. Consequently, Chomsky's claim can neither be reduced to, nor whose full content is given by, claims about a speaker's dispositions to attest to the well-formedness of given strings or, indeed, any claims about a speaker's behavioural dispositions. On this view, G rather than G' is the grammar the individual really knows even-though G and G' might have identical behavioural consequences. Thus, evidential support for G over G' would then have to be theoretical — though not for that reason any less empirical — Chomsky would urge. For instance, a theory of language acquisition might lead to the preference for one grammar G over an extensionally equivalent grammar G', on the ground that the G provides a schematism that makes available fewer grammars to the language learner.[note 7]Or a theory of "Universal Grammar" the leads to a well-confirmed grammars in a wide range of cases (i.e., enough to satisfy the predicate 'Universal') might select G over G' on the basis of available evidence (Chomsky's example). Ergo, even though a supplemented competence-theory may have behavioural consequences, no such data by themselves need lead to choice of that theory over some "extensionally equivalent" system. In sum, competence theories doubly distance themselves from behavioural data:
- from above: because competence theories do not — by themselves yield any facts about behaviour; and,
- from below: because competence hypotheses fail to be warranted solely by facts about behaviour.
para:Quine insists that such distancing from below diminishes a linguistic theory's explanatory content. "What matters," according to Quine, "is just the insistence upon couching all criteria in observation terms...[viz.] terms that can be taught by ostension, and whose application in each particular case can therefore be checked intersubjectively" [20]; "observation terms" solely pertain to a speaker's behaviour, so the question of which of two "extensionally equivalent" grammars is really the speaker's — obscures our sense[note 8]According to him, "If it is to make any [*sense] to say that a native was [*implicitly guided] by one system of rules and not by another extensionally equivalent system, this sense must link up somehow with the native's dispositions to behave in observable ways in observable circumstances"(MRCLT 444). It will not to do — to focus on dispositions to affirm the well formedness of strings, since extensionally-equivalent rules are indistinguishable on that score; different dispositions are required, if Chomsky's account is to make any sense. Quine proposes that "It could be a question of dispositions to make or accept certain transformations and not others; or certain inferences and not others" (loc. cit.). But it must(*) be a matter of some set or other of behavioural dispositions — "conjectures or conclusions... eventually be made sense of in terms of external observation" (LP 58): this is Quine's demand. His demand is not met by accounts postulating an abstract "cognitive state" that mediates between experience and behaviour — a necessary postulation, according to Chomsky, since "An attempt...to study directly the relation of behaviour to past and current experience is doomed to triviality and scientific insignificance" (RL 17). Chomsky's double departure portends twofold trouble for Quine, for it contemns his call for "clarification of criteria" (MRCLT 448).
para: Conjectures about linguistic competence tout court do not yield behavioural consequences, nor do they lead to conclusions about the hardware whose functioning such abstractly described states characterise. Additionally and crucially, even if an account of "competence" were supplemented by an account of mechanism that, given stimulus and cognitive state, yielded behaviour, Quine would remain sceptical. In Quine's view, unless there is a difference in dispositions to behaviour predicted by an account incorporating competence theory T and an account differing only through incorporation of competence theory T', there is no sense in asking which account is correct. There is a, as Quine maintains, "crying need" in linguistics "for explicitness of criteria" (MRCLT 446); and to whatever extent that "need" is not met, linguistic theory will be undesirable, since it would contain a component employing terms whose conditions of correct application are very far, if not totally, removed from observation of behavioural phenomena, which could not easily be taught by ostension, and thus terms whose applications are not guaranteed to be intersubjectively checkable.
Quine does not see black or white, but shades of grey. "Among hidden variables [*e.g., cognitive entities] there is a premium on any links to observation, however partial and indirect; the less partial the better, and the more direct the better."[21]Although some competence theories may form part of significant scientific explanations, in Quine's view the accounts involving the kind and degree of abstraction described above (i.e., Chomsky's) will not.
For Chomsky, this intentionally indirect account (of nonintensional data) helps one see the forest for the neurons; for Quine, it is a theory born of "[despair] of adhering to the standards of natural science in coping with the complexities of intelligent discourse."[22]We must deny faith in order to make room for knowledge and, in order to not be tempted by such pseudo-explanations, should "shortcut the mental bit"[23], keep our noses to the neurons, and continue isolating those dispositions to verbal behaviour most amenable to physiological explanation.
Ricketts claims that his "final characterisation of the difference between Quine and Chomsky [*amounts to cogent arguments about theory] versus persuasive rhetoric and disagreement over data (*italics are my emphasis), is suffused with a sense of dejà vu.[24]If this is true, little understanding is gained from characterising this dispute as involving different evaluations of the explanatory worth of certain theoretical approaches to "the problem of the 'causation of behaviour'." Does this interpretation allow for substantive debate rather than mere "persuasive rhetoric"?—That it does not — may be thought to follow from Quine's claim that we must work from within current science, there is no fundamental philosophical perspective which can resolve these types of issues. Verily, our conception of explanatory theory isn't the result of armchair reflection but of an inquiry into scientific toil — which is itself part of that toil. Since there is no "prior philosophy",[25] we can proceed only by seizing upon traits of the science of our day".[26]
Conclusion para: Chomsky and Quine are playing the same game in virtue of their common data, their commitment to innate structures, and their commendation of the neurological level as the level of "deepest explanation"; they have adopted different strategies because of conflicting (a posteriori) conceptions of explanation. Who wins, as both have stressed, is an empirical matter.
Source 2 (Primary):"Reply to Chomsky" (hereafter RTC) [27][edit]
Quine from the outset (of the article "RTC") expresses his belief that Chomsky was talking past him,
"Chomsky's remarks leave me with feelings at once of reassurance and frustration. What I find reassuring is that he nowhere clearly disagrees with my position. What I find frustrating is that he expresses much disagreement with what [*he thinks to be my position]." — (RTC 274), W.V. Quine |
This demonstrates concordance with Alexander George's interpretation (from Source 1) of the dispute i.e., that both Chomsky and Quine are in agreement over fundamental aspects of linguistic inquiry: what was to be considered data and what the deepest explanation would involve.
§1[edit]
Chomsky interprets Quine as believing that — linguistics is distinguished from physics since it can at best have a "tentative [*scientific] theory" for "inductive cases" and not for more complex "hypothesis formation or theory reconstruction".[28]Quine adheres to a distinction between ordinary inductive uncertainty and the deeper matter of indeterminacy of translation. However, the uncertainty familiar to physics is not the same as the indeterminacy which affects linguistics. Physics, unlike linguistics, does not rely on a "first philosophy" (RTC 275): Chomsky's "linguistic competence theory" and his "Universal Grammar" (See source 1) both arise from a priori speculation that is independent from —"above" and "below" (ibid) —the behavioural data —"the causation of behaviour"(ibid) —it seeks to explain. Moreover, the physicists appreciation of the "partial arbitrariness" (RTC 275) or underdetermination of our overall theory of nature "is not a higher-level intuition" (loc. cit.) ; it is integral to our theory of nature itself, and of ourselves as natural objects.[note 9]Contrastingly, linguistics is not just underdetermined but also indeterminate; the indeterminacy of translation is not just inherited as a special case of underdetermination it is: subtractively caliginous —an ineluctable integument that can neither be shirked nor shod.
§2: "Learning Sentences"[edit]
The principle of charity:= The more absurd the doctrine ascribed to someone, caeteris paribus, the less likely we have interpreted their words plausibly.
para: The title of this section refers to Chomsky's uncharitable interpretation of Quine, which leads to an attempt by Chomsky to provide a kind of reductio ad absurdum of Quine's views on linguistics. Chomsky attributes to Quine the absurd belief that the sentences in man's repertoire are finite in number and are generally learned as unstructured wholes. Chomsky ignores the fact that it is generally appreciated (by Quine and the scientific consensus) that the "generative" nature of grammar is what "mainly distinguishes human language from nonhuman communication systems" (RTC 276; loc. cit.). Examples of Chomsky's uncharitable nescience:
- Chomsky even quotes the following from Quine: "Our grammarian's attempted recursive specification ... will follow the orthodox line, we may suppose, of listing 'morphemes' and describing constructions."[29]
- Withal in Word and Object there are such passages as "the infinite totality of sentences of any given speaker's language" (WO 27).
- Chomsky even notices a passage that would seem to preclude his attribution (WO 71) but he unliably refuses to be swerved by it, from his systematic misinterpretations.
Quine believes that "the nature of Chomsky's misunderstanding is hinted at" (RTC 277) in the following passage:
"It ... is clear that when we learn a language we are not 'learning sentences'....Rather, we somehow develop certain principles ... that determine the form and meaning of indefinitely many sentences. A description of knowledge of language ... as an associative net constructed by conditioned response is in sharp conflict with whatever evidence we have about these matters." — (QEA), Noam Chomsky |
However, "This sense of conflict is wrong," —Quine writes (RTC 277). This "sense of conflict" comes from misinterpreting Quine as believing that language learning — is so holistic — that it involves "learning sentences outright as unstructured wholes" (RTC loc. cit.), and taking "associative net" and "conditioned response" to refer narrowly to the association of sentences as an association between unstructured wholes (italics my emphasis). Naturally, albeit ironically, Chomsky draws the consequence thusly "As far as 'learning of sentences' is concerned, the entire notion seems almost unintelligible" (QAE); if Chomsky had understood it, he would have seen that it was in no "conflict" with the doctrine that he sets over and against it, in the quoted passage. Quine writes, which Chomsky even cites, that,
"...our learning of sentences... in two modes: (1) learning sentences as wholes by a direct conditioning of them to appropriate non-verbal stimulations, and (2) producing further sentences...by analogical substitution." — WO (9), W.V. Quine |
A page later Quine adds that,
"mode (2) above is already, in a way, an association of sentences with sentences; but only in too restrained a way." — WO (10), W.V. Quine |
These passages, the majority of which Chomsky was aware of, do not lend support to Chomsky's argument...they cannot "be reconciled with the idea that I intended my phrases 'learning of sentences' and 'association of sentences' to relate to sentences [*only] as unstructured wholes" (RTC 277).
§3: "Innate Ideas"[edit]
para: Quine affirms his "penchant" for innate ideas viz. "if we construe 'innate ideas' in terms of innate dispositions to overt behaviour" (RTC 278, sic.) and this is a "penchant" he shares with behaviourists generally. A milestone in the history of empiricism was achieved when Tooke and Bentham began the "serious externalization of empiricism" (loc. cit.), which involved a shift of focus from ideas, which are private and subjective, to language which is public and intersubjective. Language aptitude is rightly characterised as innate; "language learning [*...] in which aptitude is put to work, turns on intersubjectively observable features of human behaviour and its environing circumstances, there being no innate language and no telepathy" (loc. cit.).
para: Induction, habit formation, and conditioning — essential processes; all three, necessitate discerning degrees of similarity between stimulations — innate disposition, quality space. Behavioural tests of differential conditioning and extinction of responses can be used to map a subjects quality space — to separate the innate features from the acquired ones. The points\members of a quality space are stimulations (*my emphasis), with no presupposition of dimensions. Any irrelevant features of the stimulations should disappear in the course of the experimental determination of the quality space. The actual dimensions of someone's quality space, would be obtained from ordinal comparisons of distance from the prior behavioural tests (*see sentence 2 of this paragraph); dimensionality, would be settled by pragmatic considerations of "neatest accommodation" (RTC 279).
Source 3: "Rationality in Epistemology Naturalized" (hereafter, REIN) by Edward P. Stabler, JR.[30][edit]
para: [*One of the principal] failures of the Logical Empiricist project was its program of rational reconstructionism — a form of explication which sought to replace a psychological understanding of the links between experience and science — with an understanding involving the precise tools of modern logic. *This approach failed... tout court. *If we aim to — understand, define, explain, or explicate 'intelligence' — then we must work from within science(within Neurath's boat)[note 10], and not from some illusory external vantage-point (i.e. any philosophical perspective); and that would involve, in the case of intelligence, utilising psychology and not "fabricating some fictitious structure to a similar effect" (EN 1969, p. 78)[note 11]Intelligence, as an object of knowledge, should "fall into place as a chapter of psychology and hence of natural science" (ibid p.82). The naturalist's project seeks to explain why a scientist would adopt a theoretical item such as 'intelligence'; or equivalently, "why any scientist adopts the theory he does" (ibid p. 83). *Chomsky's approach, to linguistics and cognitive science, relies on a science-external perspective to adjudicate disputes [note 12] —but there isnt such a perspective; we ought to focus, instead, on how scientific disputes are actually settled —from within science...what we will find, is that dispute-resolution is still rational...and more generally, disputes and decisions are "where rational, pragmatic" (1951, p.46).
"Section II: Departing from the Positivist Tradition." [*or, Pseudo-Rationality Confuted[note 13]][edit]
para: Carnap believed in linguistic/conceptual frameworks: he believed that empirical claims ought to be distinguished from a scientists commitment to a linguistic framework. Rational dispute can only take place when a common framework is shared, since it will have the same logical consequence and confirmation relations. However, if two disputants are using different languages, and hence different logical and confirmation relations, then they can only express their differences in a substantial way by adopting some common metalanguage (REIN 68).
para: Quine's criticisms of analyticity are engineered to show the inapplicability of this approach to scientific disputes viz., if we cannot discern "analytic truths" from empirical claims, then we will not be able to discern substantial disputes from verbal quibbles —which we must do if we want to assess a scientist's rationality. In addition, Quine's arguments for the "indeterminacy of translation" spotlight that there is —really no —objective way to decide what a scientist's linguistic framework is, since all the pertinent evidence will fail to decide among mutually incompatible theories, of the scientist's framework. Ergo, we cannot get any appropriate "criterion of analyticity", and for these reasons we cannot use Carnap's methods to assess rationality. In sum, Carnap maintained that we cannot have a rational dispute without a common framework; Quine maintains that there is no objective way of deciding whether this condition has been satisfied (REIN 60).
para:
upbraided
Intelligence and the mental[edit]
- The word 'intelligence' is largely a mentalistic term/predicate.
- Following the doctrine of anomalous monism the mental aspects of intelligence e.g., the cognitive aspects are: (a) irreducible to the physical, and (b) identical with the physical viz., with physical objects.
- Moreover, mental events make up tokens (viz., particular: beliefs) but these "tokens" do not correspond to any physical types, nor do they imply that there is any such thing as "physical types", rather, mental tokens correspond to physical tokens (e.g., neural objects). Mental particulars, correspond to particular neural objects that can't be categorised as belonging to a type i.e., wildly different areas of the brain may correspond to the same mental token.
- In addition, the mental is said to supervene upon the physical: there is no difference in the mental without a corresponding difference in the physical.
- The mentalistic aspects of 'intelligence', once explicated for science, can be viewed in one of two ways: (1) we have dispensed with them in favour of bodily terms, or (2) we have explained mental events as bodily ones; the former hostile, the latter friendly...the matter can be phrased either way.
- The mind/body problem retains its dualism, but it is a dualism of language or concepts, and not a one of ontology. Indeed, the manner in which mentalistic language classifies things, can be said to be incommensurable with the way in which physicalist language classifies things.
- Mentalistic language, however, retain its pragmatic utility, albeit as a folk-psychological theory, for our everyday interactions - where it may even be said...to be "indispensable".
- Explicating the concept of 'intelligence', involves considering whether or not it actually exists, viz. whether or not it actually refers to anything.
- Thus, in summary: if 'intelligence' refers to anything - it must be physical.
- Physicalism (as a methodological thesis) := "The unity, intelligibility, and objectivity of science rests on statements in a language of public things, events and processes in space time"[31]
- Physicalism (as an ontological thesis) := All entities, events and processes — are in space-time, and include not just matter but also energy.
Intelligence and ontology[edit]
- Our ontology consists solely of physical objects viz., objects located in space-time.
- Existence is what the existential quantifiers of FOL express, they tell us how existence is expressed i.e., via. any variables which are bound to them. The existential quantifiers range over the whole of our theory.
- Regimentation into - classical bivalent FOL (with identity) - involves Russell's theory of descriptions and the standard formation rules which generate syntax.
- 'Intelligence' can only be paraphrased as an indefinite description, such that:
- x is intelligence; AND
- x is physical
which is expressed in FOL as:
- ∃x(Ix · Px)
...there is at least one thing, such that it is intelligence and it is physical. ...where '∃' is the existential quantifier; 'I' is a singulary predicate that denotes intelligence; and 'P' is a singulary predicate that denotes physical objects; and '·' represents conjunction.
- The indefinite description of 'intelligence' represents the most general formal features, required to express the existence of 'intelligence'. When it comes to the actual study of intelligence, of course, the regimented forms will be more complex i.e., involving more logical objects e.g. connectives, and more non-logical objects viz., predicates.
- However, the subsequent "regimentation" only shows how to express the claim of existence - regarding intelligence. In order to answer the question: which existence-claims should we accept? we utilise confirmation holism and semantic holism viz., we mobilise 'holism' to assess whether 'intelligence' plays a role in supporting the meaning or confirmation of our best theories viz., taken together: if it is required as the value of an existential quantifier, that ranges over our best theories.
'Global Structuralism'[edit]
- The previous section concerned the ontology of intelligence, yet, even if there is an object that 'intelligence' refers to, what's relevant is the position it occupies in the structure of science, albeit an empty position. After all, we only know and understand the terms of a language, through their structural roles in said language, and not through the purported objects they are said to refer to.
- There isn't one "correct" explication of the explicandum 'intelligence', rather, there can be many different explications which are equally good. The structure remains undisturbed by the different candidate explications, this is due to the existence of proxy-functions.
- Proxy-functions: enable the 1-1 mapping of the objects belonging to one ontology - onto the objects belonging to another ontology viz., there exists an isomorphism between the objects of different explications.
- Even the rare possibility of underdetermination - which on the surface is unamenable to the proxy-functions - poses no threat to the structure of science: we can switch back and forth between the two incommensurable theories for richer perspectives on reality.
- What we have with structuralism is the choice among indifferent alternatives.
- Structuralism finds its canonical expression in the Ramsey-sentence, which allows us to simply be content that we are operating somewhere in the ontology of [the class of objects we have an interest in].
- There is, however, no evidence for one ontology over another viz., we accept a physicalist ontology - as opposed to say...a phenomenalist one - simply because, physical objects are vastly more efficient at organising neural input. Thus, when explicating what, and if, the explicandum 'intelligence' refers to anything, we focus on physical objects only.
- Just as there is no evidence for one ontology over another, so-too can there be no evidence for one explication of 'intelligence' over another; there are a multitude of equivalent ways for assigning values to our variables; a plethora of commensurate interpretations of the explicandum. Underdetermination itself...! proves dissoluble, instead, it unveils a more opulent perspective - on the empyrean wilderness - we call reality. The explicandum 'intelligence' may be - like all explicanda - amphibolic in nature, but much subtlety is crystallised within its ordinary use - in natural languages, it would be foolish indeed, to neglect this domain in favour of artificial languages.
Empirical research[edit]
Source 1: "Handbook of Intelligence"[32][edit]
Historical context[edit]
Section 4: "Thorndike and Connectionism"[33][edit]
para Greenwood: "He (Edward Thorndike) was not impressed by anecdotal reports of the apparently intelligent behaviour of animals, such as the apparent rationality of animals reported by Darwin[note 14]or their apparent knowledge of mathematical and mechanical principles reported by George Romanes[note 15]. Thorndike disparaged anecdotal accounts of animal behaviour, because he maintained they were generally unrepresentative of animals' cognitive abilities: he reflects on the higher probability of unexceptional animal behaviour - that is ignored in favour of cherry-picking the data - for exceptional behaviour; one example he provides is the following: "Dogs get lost hundreds of times and no one ever notices it or sends an account of it to a scientific magazine. But let one find his way from Brooklyn to Yonkers and the fact immediately becomes a circulating anecdote." another example he provides regards cats: "Thousands of cats on thousands of occasions sit helplessly yowling, and no one takes thought of it or writes to his friend, the professor; but let one cat claw at the knob of a door supposedly as a signal to be let out, and straightway this cat becomes the representative of the cat mind in all the books...In short, the anecdotes give really the abnormal or supernormal psychology of animals." sic. (Thorndike 1911, pp. 23-25).[34]
Thorndike insisted that the scientific study of animal "intelligence" should be based upon carefully controlled experimental studies. At Columbia University, Thorndike began a series of experiments in which food-deprived cats learned to escape from specially constructed cages or "puzzle boxes" to gain a food reward.[note 16][35]Thorndike's studies were based upon Lloyd Morgan's (1894)[36]explanation of how his dog Toby had learned to lift the latch on the gate of the back courtyard of his house to escape into the street. Based upon his repeated (*my emphasis) observation of the dog's behaviour, Lloyd Morgan dismissed Romanes's (1882)[37] explanation of such behaviour in terms of the dog's understanding of mechanical principles, and noted how after accidentally stumbling on a means of lifting the latch by a movement of its head, the dog had managed to learn after repeated trial and error how to open the gate. Similarly, Thorndike's cats initially clawed at the bars and pushed their paws between them, until they accidently hit on the movement required to release the latch. Like Morgan, Thorndike found that the animals took progressively less time to hit-on the required behaviour over a series of trials, until eventually they produced the learned behaviour the moment they were placed in the box.
On the basis of these experiments, Thorndike articulated what he called the "law of effect," linking learning with reinforcement, and the "law of exercise," linking learning with repetition. Thorndike was at pains to insist that learning was not based upon imitation or insight and that the connections between behaviour and response were "stamped in" by reinforcement and repetition. Consequently he called his theory "connectionism"*(which is more properly named "associationism", since "connectionism" is now associated with artificial neural networks.) Thorndike was later forced to modify the law of effect and drop the law of exercise. However, Thorndike continued to insist that learning was based upon connections or associations. So although he rejected — Darwin and Romanes's attribution of higher cognitive processes — such as rationality and mechanical understanding — to animals, he maintained that higher cognitive processes are nothing more than complex forms of connection and association, based ultimately upon the same psychological and physiological principles: "...the higher forms of intellectual operations are identical with mere association or connection forming, depending upon the same sort of physiological connections but requiring many more of them"[38].
Consequently, he supposed that an individual's level of intelligence was determined by the number of connections that individual was capable of making (*Gauss and Einstein's brains may lend support to this theory i.e. neural convolution seems to be positively correlated with intelligence ). While he recognised the difficulty of determining the nature and number of connections that needed to be made, in order to execute a complex human-cognitive-process such as language comprehension, Thorndike still went on on to develop his own intelligence tests. *Thorndike ought to be viewed, in retrospect, as a controversial figure: his alignment with racialism, and sexism ought to be unequivocally condemned!*
Qte Greenwood "Conclusion: So What is Intelligence?"(end of section 9)[note 17]:
para Greenwood: Beyond the Sturm und Drang over the inheritance of intelligence and "feeblemindedness", what was learned about the nature of intelligence itself? Not much it seemed. In a 1921 study published in the Journal of Educational Psychology that canvassed the definitions of intelligence by 14 "experts" (Thorndike et al. 1921)[39] there was remarkably little agreement, and many practitioners of intelligence testing rested content with Boring's(1923b)[40] operational definition of 'intelligence' as "what intelligence tests measure", which seemed sufficient to guaruntee its reference while saying nothing about its nature. Nevertheless, one can divine a common theme: Wundt's "apperceptive in synthesis of relational elements"; James's "discernment of similarities"; Spearman's "education of relations and correlates"; and lastly, Thorndike's "formation of connections"; they all point to some cognitive achievement involving the discernment or determination of connections and configurations." end para.
Intelligence in Nonprimates[41][edit]
[key Quote r.e. Zentall's proposal for an explicandum of 'intelligence': qte "Intelligence, in its simplest form, can be thought of as the flexibility endowed by our genes that allows organisms to adjust their behaviour to relatively rapidly changing environments."]
Notes[edit]
- ↑ The inadequately formed idea or expression is called the explicandum
- ↑ i.e. partial agreement, as opposed to synonymy, between the explicandum and its explicatum
- ↑ See his book "Wondering Significance"(2012).
- ↑ I'm aware that this review is 10 years old - obviously a limitation
- ↑ *Chomsky's 'abstract terms' are disconnected from the empirical evidence. Also, the 'abstract terms' are non-extensional
- ↑ *Again Chomsky has not provided us with any clear — and explicit methodology — or criteria.
- ↑ G and G' are extensionally equivalent, because they weakly generate the same set of strings: their outputs (the speech they produce) are identical viz., under certain weak constraints.
- ↑ My use of the word "sense" means: our understanding or ways of understanding something.
- ↑ Quine might reconfigure the ancient Greek maxim "know thyself" as "Nature know thyself". See Alex Orenstein's entry "Quine and the Problems of Consciousness" in Consciousness and the Great Philosophers.
- ↑ See this wikipedia article for an overview of the Neurath's similie: Neurath's Boat
- ↑ It is understandable [...] that the philosopher should seek a transcendental vantage point, outside the world that imprisons [the] natural scientist and mathematician. He would make himself independent of the conceptual scheme which it is task to study and revise. "Give me που στω [a place to stand]," Archimedes said, "and I will move the world." However, there is no such cosmic exile. [...] The philosopher is in the position rather, as Neurath says, "of a mariner who must rebuild his ship on the open sea." —W.V. Quine (notes for Sign and Object, Nov 5th 1944) [*I found this quote as an epigraph in the book "Working from Within" by Sanders Verhaegh (2018)]
- ↑ See the penultimate paragraph of section 1 for more details
- ↑ [* What is pseudo-rationality? I am indebted to Otto Neurath for coining the term; pseudo-rationality, principally, involves not acknowledging the limitations that are inherent to rationality. See this wikipedia article for what seems to be a decent overview: Pseudorationality.
- ↑ Reported in The Descent of Man, and Selection in Relation to Sex; wherein, Darwin claimed that most people would agree that "animals posses some power of reasoning," on the grounds that they "may constantly be seen to pause, deliberate, and resolve" (1871, p. 46).
- ↑ Reported in Animal Intelligence(1882)*page number unreferenced
- ↑ This was part of his doctoral dissertation of 1898, and published later that year as a monograph supplement in Psychological Review
- ↑ I have only included Thorndike but "section 9" also included: William James, Wilhelm Wundt, Alfred Binet, Spearman,Yerkes, Goddard, and Cattell, viz., it provided a survey of definitions of 'intelligence'.
References[edit]
- ↑ Deary, Ian J., Lars Penke, and Wendy Johnson. "The neuroscience of human intelligence differences." Nature reviews neuroscience 11.3 (2010): 201-202;
- ↑ Quine, W.V., 1953. The problem of meaning in linguistics (p. 47-48)
- ↑ Harris, Zellig S., and Donald C. Swanson. Language, vol. 16, no. 3, 1940, pp. 216–235. JSTOR, www.jstor.org/stable/409060.
- ↑ George, A., 1986. Whence and whither the debate between Quine and Chomsky? The Journal of philosophy, 83(9), pp.489-499.
- ↑ George, A., 1986. Whence and whither the debate between Quine and Chomsky?. The Journal of philosophy, 83(9), p. 492; (henceforth I will simply state the relevant page number instead of the entirety of the citation.
- ↑ Ontological Relativity and Other Essays (New York: Columbia, 1969), p. 27. Quine, W.V.
- ↑ Putnam, Hilary, 'The "Innateness Hypothesis" and Explanatory Models in Linguistics', Synthese 17 (1967) 12-22.
- ↑ Bechtel, P.W., 1978. Indeterminacy and intentionality: Quine's purported elimination of propositions. p. 654.
- ↑ "States of Mind," this JOURNAL, LXXXII, 1 (January 19885): 5 - 8, p. 6.
- ↑ "Some Conceptual Shifts in The Study of Language," in L. Cauman, I. Levi, C. Parsons, and R. Schwartz, eds., How Many Questions?: Essays in Honour of Sidney Morgenbesser (Indianapolis: Hackett, 1983), p. 157.
- ↑ "Changing Perspectives on Knowledge and Use of Language," 1984, unpublished manuscript, p. 10.
- ↑ Reflections on Language (RL hereafter) (New York: Pantheon, 1975), p. 24.
- ↑ Ricketts, T.G., 1982. Rationality, translation, and epistemology naturalized. The Journal of Philosophy, 79(3), pp.117-136. p.136
- ↑ Chomsky, Noam. "Rules and representations." (1980): 1-15.
- ↑ "Changing Perspectives on Knowledge and Use of Language," 1984, unpublished manuscript, p. 10. Chomsky, Noam
- ↑ Quine, W.V., 1972. Methodological reflections on current linguistic theory. In Semantics of natural language (pp. 442-454). Springer, Dordrecht.
- ↑ Chomsky, N., 1969. Quine’s empirical assumptions. In Words and objections (pp. 53-68). Springer, Dordrecht.
- ↑ Chomsky, N., 1965. Aspects of the Theory of Syntax. (MIT press), p. 8
- ↑ Quine, W.V., 1972. Methodological reflections on current linguistic theory. In Semantics of natural language (p. 446). Springer, Dordrecht.
- ↑ Quine, W.V.O., 1969. "Linguistics and philosophy", p. 58
- ↑ Quine, W.V., 1980. "Sellars on behaviorism, language, and meaning". Pacific Philosophical Quarterly, LXI, 1 (Jan 1980): pp. 26-30.
- ↑ Quine, W.V., 1975. "Mind and verbal dispositions" (hereafter MVD). p. 86.
- ↑ Quine, W.V., 1985. "States of mind" (hereafter SM). The Journal of Philosophy, 82(1), p. 6.
- ↑ Ricketts, T.G., 1982. Rationality, translation, and epistemology naturalized (hereafter RTEN). The Journal of Philosophy, 79(3), p. 136.
- ↑ Quine, W.V.O., 1969. Ontological relativity (hereafter OR). Ontological relativity and other essays, 26, p.68.
- ↑ Quine, W.V., 1957. The scope and language of science (hereafter SLS). The British Journal for the philosophy of Science, 8(29), p. 244.
- ↑ Quine, W.V., 1969. "Reply to Chomsky". Synthese (1968). pp. 274 - 283
- ↑ Chomsky, N., 1969. Quine’s empirical assumptions (hereafter QEA). In Words and objections [loc. cit.] [original source is Synthese] [pp. I'll add later].
- ↑ Quine, W.V., 1961. "From a logical point of view" (hereafter FLPV): Nine logico-philosophical essays (No. 566). Harvard University Press.
- ↑ Stabler Jr, E.P., 1984. Rationality in naturalized epistemology. Philosophy of science, 51(1), pp.64-78.(Received June 1983; Revised October 1983).
- ↑ 1931a/1983, "Physicalism: The Philosophy of the Vienna Circle", pp. 48-51
- ↑ Goldstein, S., Princiotta, D., & Naglieri, J. A. (2015). Handbook of intelligence. Evolutionary theory, historical perspective, and current concepts.
- ↑ Section 9: Intelligence Defined: Thorndike and Connectionism. Greenwood, John D. pp. 127-128
- ↑ Thorndike, E.L.(1911). Animal Intelligence. New York: Hafner.
- ↑ Animal Intelligence: An Experimental Study of the Associative Process in Animals(Thorndike 1898)
- ↑ Morgan, C. L. (1894). Introduction to comparative psychology. London: Walter Scott
- ↑ Romanes, G. J. (1882). Animal Intelligence. New York: Appleton.
- ↑ The measurement of intelligence. New York: Teachers College Press. Thorndike et al. (1926), p. 415
- ↑ Thorndike, E.L., et al. (1921). Intelligence and its measurement: A Symposium. Baltimore.
- ↑ Boring, E.G. (1923b). Intelligence as the tests test it. New Republic , 3, 35-37.
- ↑ §2: "Intelligence in Nonprimates", Zentall, Thomas R.