To be presented at the Wittgenstein Colloquium
WITTGENSTEIN’S
VERIFICATIONISM AND THE GREATEST BLUNDER OF ANALYTIC PHILOSOPHY
Summary
This
paper reconsiders the legacy of verificationism by returning to its original
source in Wittgenstein’s thought. The logical positivists, in their attempt to
deploy Wittgenstein’s insight as a weapon against metaphysics, misconstrued his
project by recasting it into rigid logico-semantic formulations. In doing so,
they constructed a strawman version of verificationism – one that ultimately
proved untenable and was subsequently abandoned. The academic influence of
their successors ensured that this rejection hardened into “received wisdom”:
verificationism came to be regarded as a philosophical dead end.
I
argue that this conclusion represents a profound misstep in contemporary
analytic philosophy. To move beyond it, we must disentangle Wittgenstein’s
original insight from the positivist distortions. What emerges is not a failed
semantic theory, but rather a viable framework for a pragmatic investigation of
representative language. Wittgenstein’s verificationist idea, properly
understood, illuminates the distinction between cognitive meaning and its
relation to truth-values, while opening the way to a broader pragmatic analysis
of how representative language functions.
The
paper proceeds by offering a series of concise responses that a Wittgensteinian
verificationist might give to some of the principal objections traditionally
raised against the view. In doing so, it seeks to rehabilitate verificationism
as a workable and philosophically fruitful approach, once freed from the
misinterpretations of logical positivism.
Keywords:
Wittgenstein,
verificationism, cognitive meaning.
My thesis on Wittgensteinian
verificationism is the following: As it is well known, he was the creator of
the idea (Cf. Glock: 354). The philosophers of logical positivism appropriated
his insight as a tool for their attack on metaphysics, attempting to develop it
into a general logical-semantic principle. In doing so, they created a rigid ristraw man that had little to do with the
principle Wittgenstein had originally proposed to them. However, they soon
realized that this straw man could not stand, concluding that the verification principle
was untenable. Since they were influential, this conclusion was passed down as inherited
wisdom. As a result, verificationism seems today dead and buried. However,
when we turn to a careful consideration of the principle as proposed by Wittgenstein,
we see that it remains perfectly defensible. In my judgment, this misunderstanding
is the greatest blunder of contemporary analytic philosophy. (Costa
2018, V) In what follows, I will try to show why.
1
To highlight the contrast,
I begin by presenting the formulation of verificationism initially proposed by the
logical positivist A. J. Ayer. In his words:
The mark of a genuine factual proposition
is that some experiential propositions can be deduced from its conjunction to
certain other premises without being deducible from these other premises alone”
(1952: 38–39).
Calling the
factual proposition S, the other premises P, and the observational result
deduced O, we can formulate this as: “S & P ⊢ O”.
The problem, soon recognized by Ayer
himself, was that the criterion is too liberal, rendering all propositions true.
If S is “The Absolute is lazy” and P is “If the Absolute is lazy, then snow is
white,” since this implication is true and snow is indeed white, it must then
be true that the Absolute is lazy. Moreover, it seems that S cannot be true or
false without already being meaningful.
Just like this first formulation by Ayer,
other positivist reformulations of the verification principle proved untenable,
as Carl Hempel’s article (1950) has shown.
2
I now consider how
Wittgenstein addressed the question. Here are a few quotations from 1929-1930:
The meaning of a sentence (Satz) is
its method of verification. (1984b: 29)
A sentence without a way of verification
has no sense (Sinn). (1984a: 245)
If two sentences are true or false under
the same conditions, they have the same sense. (1984a: 244)
- The method of verification is not a
means, a vehicle, but the sense itself. Determine under what conditions a sense
must be true or false, thus determine the meaning of a sentence. (1984a: 47)
For Wittgenstein,
the principle is the full cognitive meaning of a sentence, which is true if it
is applied and false if not. The method of verification is the same as the “propositional
content”, a version of what Frege called “thought” (Gedanke). Around
1932–35 Wittgenstein presented a particularly illustrative example annotated by
Alice Ambrose. It is worth quoting in full:
Consideration
of how the meaning of a sentence is explained makes clear the connection
between meaning and verification. Reading that Cambridge won the boat race,
which confirms that ‘Cambridge won,’ is obviously not the meaning, but is
connected with it. ‘Cambridge won’ isn’t the disjunction ‘I saw the race, or I
read the result, or...’ It’s more complicated. But if we exclude any of the means
to check the sentence, we change its meaning. It would be a violation of
grammatical rules if we disregarded something that always accompanied a
meaning. And if you dropped all the means of verification, it would
destroy the meaning. Of course, not every kind of check is actually used to
verify ‘Cambridge won,’ nor does any verification give the meaning. The
different checks of winning the boat race have different places in the grammar
of ‘winning the boat race.’ (2001: 29)
The principle
appears here as a variety of means or ways of verification which
together constitute the meaning of the declarative sentence. If we remove one means
of verification, we can remove part of the sentence’s meaning. If we remove all
means of verification, it ceases to have meaning.
Moreover, these means differ in value. Some are
fundamental, strongly contributing to the meaning, such as seeing the Cambridge
team’s boat win and hearing the referee’s whistle; others contribute less to
the meaning, such as hearing someone say that Cambridge won.
Furthermore, there is a causal relation
between verification by watching the race and by hearing about, reading in a
newspaper, having seen the trophy at the club... Verification is usually
branched, like a tree: the trunk being direct observation, and the branches,
dependent on the trunk. Moreover, what one means by an assertive
utterance can be aspectually emphasized, for example, when I say I know because
someone told me so.
It is possible, however, to present a more
precise general formulation of verificationism à la Wittgenstein,
according to which meaning naturally flows from the declarative sentence as ways
of verification. To this end, I consider S to be any declarative sentence and
call the verification rule the sum of means of verification hierarchically
constituting the meaning of S. As a result, the principle of verification or VP
becomes:
VP (Df.) = the cognitive meaning of a
declarative sentence S = the rule of verification for S.
Consider the analysis
of a simple example using VP regarding the sentence (i) “This piece of metal is
magnetized.” Following Wittgenstein’s advice, if we try to explain the
meaning of (i), we will find its verification rule understood as a sum of means
of verification. One explanation is as follows: “A magnetized piece of metal
attracts iron objects; this is a piece of metal; This piece of metal attracts pieces
of iron; hence (i) is true. The meaning of (i) is the verification rule, whose
effective application equals its assertion. The rule is nothing more than what
we understand or mean by (i). Nothing could be more intuitive.
My conclusion is that Wittgenstein’s
principle of verification should not be seen in terms of an essentially
logic-semantic principle, as the positivists wished, but rather as constitutive
of a pragmatics of factual discourse, showing that we should select and
organize the principal types of declarative sentences and analyse their
verification rules in an investigation reminiscent of the speech acts theory.
3
Let us now
consider some main criticisms made of the principle of verification in the
light of the formulation above (Costa 1984, V).
The first is that the principle is
self-refuting. It is either analytic or synthetic. If it is analytic, then it
must be non-informative and its negation contradictory. But since it seems to
be neither of these, it is synthetic. Yet if it is synthetic, then it must be
verifiable in order to have meaning. But when we try to apply the principle to
itself, we see that it is unverifiable. Therefore, it is devoid of meaning!
A Wittgensteinian response would be that the
principle is a general grammatical sentence about the way all of our factual
language must work so that its sentences may reach truth-values. My way of
presenting this is to say that the VP is, in fact, analytic, for all it does is
to make explicit a hidden synonymy between “meaning as the cognitive content of
a declarative sentence” and its analytic unpacking as “the rule, i.e.,
the hierarchized means by which we establish the truth-value of its cognitive
content.”
Analytic sentences of hidden synonymy are common.
Consider the sentence “6514 = 3,257 + 3257.” This is an identity sentence, but
only a savant would perceive that it is just as analytic as “4 = 2 + 2” or
that its negation is contradictory. Therefore, we have good reasons to admit
that VP is an analytic principle.
4
Let us now
consider W. V-O. Quine’s objection. It was inspired by Duhem’s holism,
according to which it is impossible to confirm a scientific hypothesis apart
from the constitutive assumptions of the theory to which it belongs. In Quine’s
concise statement: “…our statements about the external world face the tribunal
of sense experience not individually but only as a corporate body.” (1951: 9)
From this followed his semantic holism, according to which our language
forms a network of meanings that cannot be divided up into verifiability
procedures explanatory of the meaning of isolated statements.
The problem with Quine’s holism is that he
seems not to have noticed that the statements of a theory are not all
verified at once. The constitutive assumptions of the linguistic field to which
the statement belongs are first and separately verified. Finally, the statement
we have in mind is verified under the presupposition that those statements have
already been verified.
Consider Galileo’s statement: “Jupiter has
four moons.” He verified it by telescopic observation, night after night, of
four luminous points to the right and left of Jupiter that systematically moved
from one side of the planet to the other. His obvious conclusion was that they
were moons of Jupiter. But there were constitutive assumptions that had already
been verified, such as the law of the telescope, according to which the power
of magnification results from the focal length of the telescope divided by the
focal length of the eyepiece, that Jupiter is a planet orbiting the Sun like
the Earth, that the Moon is Earth’s satellite, etc. It is obvious that
Galilei’s way of verification of Jupiter’s moons comes after all these already
verified constitutive assumptions, being therefore detached from them in full correspondence
with the meaning of the statement.
5
I now want to
consider the objection of existential-universal asymmetry. The idea is that I
can verify “This piece of copper expands when heated”, but I cannot verify “All
pieces of copper expand when heated”, since in order to do so, I would have to
observe the heating of all pieces of copper in the world. This finding makes
the universal laws of physics unverifiable.
The answer lies in the distinction between
what is absolutely certain and what is practically certain, and
in the false belief that in order to be conclusive, one must be absolutely
certain. Consider the sentence “I affirm that 2 + 2 = 4.” This sentence is
absolutely certain. Now consider: “I affirm that all pieces of copper expand
when heated.” This sentence is practically certain. It has been sufficiently verified
to be beyond doubt. Of course, it remains probabilistic. Even so, it is
conclusive. The same with the laws of the universe.
6
Let us now
consider the issue of arbitrary indirectness. Consider the sentence “The mass
of an electron is 9.109 x 10 kg raised to the thirty-first negative power.’
Cases like this force us to admit that many verifiability rules are based on indirect
observation. But how can we distinguish direct from indirect observation? Is
this not a desperately confusing distinction? (Lycan 2000: 121).
To
this, one could answer that our assertoric sentences are always made within
what Wittgenstein called practices, language games, linguistic
regions. Consequently, our distinction between direct and indirect
observations should always be taken against the background of a linguistic
practice, without assuming the practice of our everyday observation as the only
true one.
Thus, consider the bacteriologist’s
linguistic practice. She will say that she verifies a cell with deformations
directly in her microscope, but that she can, in this way, indirectly verify
that there are víruses infecting the cell. Consider now the archaeologist’s
linguistic praxis. They will argue they indirectly know that humans inhabited
North America 21,000–23,000 years ago because of the direct Discovery of human
footprints in New Mexico. My conclusion is that there is no problem in
distinguishing between direct and indirect verification, provided that we take
into account the verification’s linguistic practice.
7
There are also empirical
counterexamples. I consider only one, proposed by Michael Dummett (1978: 148
ff): “John was courageous”, when John died without having had any opportunity
of demonstrating courage. Assuming that the only way to verify that John was courageous
would be by observing his behavior, this sentence seems unverifiable. Hence, it
should be senseless. But it seems meaningful. The answer is that this sentence
has a grammatical meaning but no cognitive meaning. To make this clear,
suppose you go on a hike in a remote place and find written on a stone, “John
loves Mary”. This sentence has a grammatical meaning, but as you do not know
which John and Mary were, you cannot give it any cognitive meaning.
8
Now I wish to
consider two opposing cases of formal sentences that are said to be meaningful
but unverifiable. Verification here means proof based on axioms. Consider
Goldbach’s conjecture:
G: Every even number greater than 2 can be
expressed as the sum of
two prime numbers.
The objection is that this conjecture has cognitive meaning,
though no one has yet found a proof that verifies it. Our answer is: you are reading
a conjecture as if it were a theorem. As a conjecture, its true form is: “It is
plausible that G”. Indeed, G is plausible because until now all prime numbers
we have found as expressing the sum of two primes. Hence, as a conjecture, G is
true, as it has been verified.
Consider now the case of Fermat’s last
theorem:
F:
There are no three positive integers x, y, and z that
satisfy the
equation
xⁿ + yⁿ = zⁿ, if n is greater than 2.
It was only proven
in 1995 by Andrew Williams. Now, one could argue that before 1995 the theorem
existed and was meaningful, though there was no verification in sight. Our
answer is that to say that F was a theorem before 1995 was a misnomer. It was
only a conjecture. Its real form was “It is plausible that F”. In fact, the
origin of the misname was Fermat himself, since he jokingly wrote that he had a
proof of F that he couldn’t put on paper since the margins of his notebook were
too narrow.
9
The last objection
I would like to answer here comes from Quine’s rejection of the distinction
between analytic and synthetic. As I take the verification principle as
analytical, it makes sense to address this possible charge.
Quine defined an analytical sentence in a
Fregean way as tautological (true by logical constraints) or shown as
tautological by the replacement of its non-logical terms with cognitive
synonyms. However, he found the word ‘synonymous’ in need of explanation. His
first answer was that a synonym is a word that can be replaced by another in
all contexts salva veritate. However, this answer does not work in all
cases: “creature with heart” and “creature with kidney” are not synonymous, but
can be replaced in all contexts salva veritate, since their extensions
are the same.
In a further attempt to define analyticity,
Quine made an appeal to the modal notion of necessity: “Bachelors are unmarried
males” is analytic if and only if necessarily, bachelors are unmarried
males. But he also saw that the usual notion of necessity does not cover all
cases. Phrases like ‘equilateral triangle’ and ‘equiangular triangle’ necessarily have the
same extension, but are not synonyms. Consequently, we must define ‘necessary’,
in this case, as the specific necessity of analytic statements, so that the
concept can be applied in all possible circumstances... However, the ‘necessity
of analyticity’ is an obscure notion, if it really exists. Dissatisfied, Quine
concluded that any attempt to explain analyticity, if not circular, “has the
form, figuratively speaking, of a closed curve in space.” (Quine 1951: 8)
A well-known response
to Quine is that a word could not be defined by words that do not belong to its
specific field. We do not define words belonging to ornitology using words from
quantum physics and vice versa; thus, we should not try to define analyticity
by means of words like necessity.
My objection is that
Quine’s attempt took the wrong turn. It seems more sensible to take recourse to
the dictionaries’ definitions like:
Synonymous = words of the same language
that have the same or nearly the same meaning (Webster 1995)
This can be read as saying that two words are synonymous if they
have similar (identical or nearly identical) meaning-definitions that give
their meanings back.
“A creature with a heart” and “a creature with
a kidney” are not synonymous, since in the first case the definition is of an
animal that has an organ to pump the blood, while in the second is of an animal
with an organ to clean the blood. An ‘equilateral triangle’ and ‘an equiangular
triangle’ are not synonymous because the former is defined as a triangle whose
three sides are equal, while the latter is defined as a triangle whose three
internal angles are congruent with each other.
Moreover, since
meaning-definitions are made to have the same meaning as their definienda,
they are also synonymous. Hence, we can replace Quine’s flawed definition of
analyticity with a more adequate definition, expecting that the tautologies
generated by analytic statements only require replacement by their meaning-definitions.
This is the case with the sentence
“Bachelors are unmarried adult males”. Defining bachelor as “an unmarried adult
male” and replacing the subject ‘bachelor’ with its definition, we produce the tautology
“Unmarried adult males are unmarried adult males”. Of course, our natural
language is inherently and purposefully vague. What precisely is a marriage, or
a male? At what age can a male be considered an adult? Higher precision here would
demand stipulation.
We
should say the same about the definition of verification. Definiendum
and definiens are here synonymous by precise similarity, not by formal
identity. And so it is good. After all, as Aristotle wrote: “It is the mark of
an educated man to seek precision only so far as the nature of the subject
admits.” (1985: 1094b).
REFERENCES
Ayer, A. J. (1952). Language, Truth,
and Logic. New York: Dover.
Aristotle (1985). Nicomachean Ethics,
in The Complete Works of Aristotle. Ed. J. Barnes (Princeton University Press, 1985).
Costa,
Claudio (2018). Philosophical Semantics: Reintegrating Theoretical
Philosophy. Newcastle upon Tyne: Cambridge Scholars Publishing.
Dummett, Michael (1978). Truth and Other Enigmas. Cambridge
MA: Harvard University Press.
Glock, H-J (2010). Wittgenstein’s Lexicon.
Darmstadt: Wissenschaftliche Buchgesellschaft.
Harkaway & Co. (1995). Webster’s New
Encyclopedic Dictionary. New York: Black Dog & Leventhal.
Hempel, Carl (1959) “The Empiricist Criterion of
Meaning”, in A. J. Ayer (ed.) Logical Positivism. New York: Free Press.
First published under the title “Problems and Changes in the Empiricist
Criterion of Meaning”, in The review internationale de philosophie, vol.
11, pp. 41-63.
Lycan (2000). Philosophy of Language: A
Contemporary Introduction. London: Routledge.
Quine, W. V. O. “Two Dogmas of Empiricism”, Philosophical
Review 1951, 60 (1), 20-43.
Wittgenstein, Ludwig (1984a). Wittgenstein und der
Wiener Kreis. Werkausgabe Band 3, Frankfurt am Main: Suhrkamp.
Wittgenstein, Ludwig (1984b) Philosophische
Bemerkungen. Werkausgabe Band 2, Frankfurt: Suhrkamp.
Wittgenstein, Ludwig (2002). Wittgenstein’s
Lectures – Cambridge 1932-1935. Ed. Alice Ambrose, New York: Prometheus
Books.
