Selected Criticisms of Wittgenstein:

Copyright © 2014 Bruce W. Hauptli

Criticisms of the Early Wittgenstein’s views:

1. In his Pulling Up The Ladder, Richard Brockhaus offers the following criticism of the early Wittgenstein:

there are now three classes of propositions with which to deal: elementary propositions, explicit truth-functions of those elementary propositions, and unanalyzed propositions of ordinary language.  Even if we accept that the first two are pictures, we must still explain how human beings recognize the senses of unanalyzed ordinary propositions, which are not obviously pictures at all.  Without such an explanation the picture theory cannot provide an adequate explanation of...[many] linguistic competencies....[1] 

2. In his Wittgenstein’s Metaphysics, John Cook maintains that:

...Wittgenstein claimed that he had found, in all essentials, the solution to philosophical problems.  To make good this claim, however, he would have to show how philosophical disputes over logical form are to be settled.[2] 

Wittgenstein was well aware that he had not the faintest idea of how to carry out an actual analysis.  We are at present, he said, “unable to give the composition of elementary propositions.”[3]

The truth about the Tractatus, then, is this: although it argues that certain conditions would have to be met by anything that counts as a language, it fails to demonstrate that the propositions of everyday language actually fulfill these conditions.  If they were found not to be analyzable in the anticipated way, Wittgenstein would have had to conclude that they do not describe possible states of affairs and so say nothing at all.[4] 

3. The role of logical analysis and the truth-table model poses a problem in that if there are very many atomic elements, the prospects for a truth-table analysis of complex statements becomes dubious.  In his “Internalism Exposed,” Alvin Goldman asks:

as Christopher Cherniak points out, determination of even tautological consistency is a computationally complex task in the general case.  Using the truth-table method to check for consistency of a belief system with 138 independent atomic propositions, even an ideal computer working at “top speed” (checking each row of a truth table in the time it takes a light ray to traverse the diameter of a proton) would take twenty billion years, the estimated time from the “big bang” dawn of the universe to the present.  Presumably, twenty billion years is not an acceptable doxastic decision interval![5] 

Selected. Criticisms of the Later Wittgenstein’s views:

1. In his The Problem of Knowledge, Alfred Ayer offers a criticism of Wittgenstein’s discussion of “family resemblance” (regarding games and knowledge):

 this is a good analogy, but I think that Wittgenstein is wrong to infer from it that games do not have any one thing in common.  His doing so suggests that he takes the question whether things have something in common to be different from the question whether there are resemblances between them.   But surely the difference is only one of formulation. If things resemble one another sufficiently for us to find it useful to apply the same word to them, we are entitled to say, if it pleases us, that they have something in common.[6] 

2. In his The View From Nowhere, Thomas Nagel critiques (an over-extension of) Wittgenstein’s 5 o’clock on the sun discussion.[7] 

3. In his Empirical Knowledge, Alan Goldman notes that:

this Wittgensteinian notion of a criterion as necessary evidence conceptually linked to a state of affairs, or central importance to this current school of epistemologists, will be explicated and criticized much later in our story [cf., Chapter 14].  I shall argue that the notion is ultimately incoherent in itself.  There simply is no fully coherent position between a phenomenalism or idealism that does away with the physical world as we conceive it and a full-blooded realism that acknowledges the world’s independence from our efforts to know it.  For the realist, the argument against phenomenalism on linguistic grounds rejects a wrong view for the wrong reason; and the attempt to save some noncontingent conceptual connection between perceptual evidence and knowledge of objective properties retains the weakest premise of only one version of the older position.[8] 

4. In his The Fragmentation of Reason, Stephen Stich maintains that: “some writers have been tempted by the Wittgensteinian idea that epistemic assessments must come to an end with the criteria embedded in our ordinary concepts of cognitive evaluation.  But surely this is nonsense.  Both our notions of epistemic evaluation and (more important) our cognitive processes themselves can be evaluated instrumentally.”[9] 

5. In his Unnatural Doubts, Michael Williams maintains that:

...Wittgenstein shows that the concept of knowledge cannot meaningfully be applied to “hinge” or “framework” propositions.  This is what is meant by the claim that our relation to them is “non-epistemic.” 

  ...the thought that our relation to various basic certainties is non-epistemic has to be understood this way.  Otherwise, the sceptic will say that, though our relation to such judgments is ordinarily non-epistemic, things change in the context of epistemological investigation.  But he wants for force the critic to argue not just that our relation to them is not epistemic but that it cannot be....The anti-skeptical strategy in question begins to look less like a satisfactory diagnosis of scepticism than a simple reassertion of ordinary certainties against the unsettling results of philosophical reflection. 

  The problems with this “Wittgensteinian” response to sceptics do not end here.  The thought that we have a special non-epistemic relation to certain “framework judgments” suggests that these judgments, while genuinely factual, have a privileged placed in the scheme of things, and we can now see why such a view will be hard to sustain.  No matter what account we try to give of this privileged place, it will be difficult to deny the septic his triumph if we admit that these judgments are ultimately groundless, while continuing to insist that they are genuinely factual.  In consequence, anyone who goes down the path of claiming that we enjoy a special “non-epistemic” relation to our basic commonsense certainties will come under strong pressure to deny that these judgments are straightforwardly factual, which brings us back to the question of whether the envisaged reply to the sceptic offers more than verbal camouflage for a large concession.  Once more, the epistemologist’s dilemma.[10] 

They are hinges on which all empirical investigation turns.  But the difficulty is to see why this is a refutation of radical scepticism rather than another expression of it.  Suppose we give Wittgenstein the concept of knowledge: still, aren’t we explaining the restriction on its scope in terms of the impossibility of justifying certain propositions?  If this is not to amount to conceding the sceptic’s point, it must be shown that this absence of justification is not a lack.[11] 

It will not do to object that Wittgenstein shows more than that we do not treat certain hinge propositions as open to question but that, since their standing fast for us constitutes what we understand by judgment, we cannot so treat them.  At least, it will not do unless we have a way of meeting the reply that “cannot” only means “cannot if we are to get on with ordinary pursuits.”  As historians, we cannot entertain radical doubts about the reality of the past.  But this does not mean we cannot entertain them as epistemologists.[12] 

6. In his “Realism With A Human Face,” Hilary Putnam discusses the semantic paradoxes and maintains that:

but the problem is that the things which we are “shown” when Semantic Ambiguity is explained to us are shown by being said.  The idea that there are discursive thoughts which cannot be “said” is just the formalistic trick that I said I don’t understand.[13] 

7. In her Considered Judgment, Catherine Elgin maintains that:

if a sequence of behavior displays no discernible resemblance to the activities we count as games, if that behavior does not satisfy our criteria for ‘game’, we do not (and should not) call it a game.  Since the meaning is the use, it follows from our collective refusal to call it a game that it is not a game.  Our word ‘game’ does not apply to it.  Similarly, activities that do not satisfy our criteria for ‘language’, ‘practice’, and ‘form of life’ are ipso facto not languages, practices, or forms of life.  The world then could not be rife with unrecognizable intelligences, for beings would not be intelligences if we did not recognize them as such. 

  Pure procedural epistemology precludes incommensurability.  If translation is possible, languages are commensurable.  Real disagreements are resolvable, and seemingly irresolvable disagreements are, for that very reason, unreal.  So if Wittgenstein is right about language, he is wrong about lions.  If a lion could talk, we could understand him; for if we cannot understand him, it follows that he cannot talk.  If translation is impossible, the phenomenon we confront is not a language.  Our language then is comprehensive: we can be confident that it captures every possibility, for whatever is inexpressible is inconceivable; and whatever is inconceivable is impossible.  So the adequacy of language is assured.  Whatever its limits, noting is left out.[14] 

Our cognitive practices thus apparently differ from games in permitting newly discovered facts, newly developed tests, newly formulated desiderata to reinforce or discredit currently accepted findings and the standards that justified their acceptance. 

  Pure proceduralists must deny this.[15] 

Change is just prospective; correction is retrospective as well.  Though old victories stand when the rules of basketball are revised, old diagnoses may fall to advances in medical science. 

  Moreover, cognitive standards and methods, unlike rules of a game, require validation.  The IQ test is in disrepute, not because its results are irreproducible but because their relation to the magnitude they purport to measure is moot.  If an IQ score were the standard of intelligence in the way a basketball score is the standard of winning, test results would speak for themselves.  The do not.  In demanding validation, psychology concedes that it does not consider the IQ test the authoritative measure of intelligence.[16] 

Although we regularly act on our findings, that is not all we do.  We hedge our bets.  We devise safeguards to avert error and damage control mechanisms to channel the effects of errors we cannot prevent and to soften the impact.  We hone techniques, refine methods, sharpen standards, and then deploy them to evaluate the continued acceptability of previously accepted results.  Were satisfaction of current standards sufficient for rightness, such efforts would be otiose; were it constitutive of rightness, they would be unintelligible.  They are neither.  When standards function as reasonable but potentially fallible indicators of rightness, such prudence makes sense.  A distinction between being right and satisfying our standards for rightness is implicit in cognitive practice.  It derives from the design of the practice, not the quality of its product; so it involves no admission that current standards are flawed.  But in providing for the possibility that its standards are unreliable, our practice denies that they are authoritative.[17]