RUDOLF CARNAP ON SEMANTICAL SYSTEMS AND
W.V.O. QUINE'S PRAGMATIST CRITIQUE
BOOK III - Page 6
Quine's Pragmatist Critiques
Willard Van Orman Quine (1908-2000) was born in Akron, Ohio. In 1930 he graduated summa cum laude in mathematics from Oberlin College, and then entered Harvard University’s graduate school of philosophy. He wrote his doctoral dissertation under the direction of Alfred North Whitehead, the co-author with Bertrand Russell of the Principia Mathematica, and he published the dissertation as A System of Logistic in 1934. Quine became a faculty member of Harvard’s department of philosophy in 1936, where he remained for the duration of his long career. He enjoyed traveling, and wrote an autobiographical travelogue as The Time of My Life in 1985. Quine described his long acquaintanceship with Carnap in “Homage to Rudolf Carnap” (1970), a memorial article published in the year of Carnap’s death, and reprinted later in Quine’s Ways of Paradox (1976). Quine met Carnap during his European travels in the 1930’s, and their dialogues continued after Carnap relocated to the United States in 1935. Their private correspondence has been published under the title Dear Carnap, Dear Van (ed. Creath, 1990), which reveals nothing about their philosophical views that is not already known from their published works, but exhibits their enduring friendship notwithstanding their philosophical differences.
While Quine might be regarded as Carnap’s principal protagonist, their philosophies are much more similar than they are different. In the memorial article Quine refers to Carnap as a towering figure, who dominated philosophy in the 1930’s as Russell had in previous decades, and he also refers to Carnap as his greatest teacher. Quine’s best known criticism of Carnap’s philosophy is his rejection of the analytic type of statement. This criticism together with several others has their basis in Quine’s pragmatist view of empiricism. Quine published a brief statement of his own doctrine of empiricism as “The Pragmatist’s Place in Empiricism” (1975), later appearing in his Theories and Things (1981) as “Five Milestones of Empiricism.” This paper is ostensibly a history of empiricism in terms of five historical turning points, but the five historical milestones also happen to be the central theses of Quine’s own pragmatist philosophy. He summarizes these five historical turning points as follows:
1. The shift from ideas to words
2. The shift of semantic focus from terms to sentences
3. The shift of semantic focus from sentences to systems of sentences
4. The abandonment of the analytic-synthetic distinction
5. The abandonment of any first philosophy prior to natural science
Quine’s several criticisms of Carnap’s positivist version of empiricism may be viewed as having a basis in these five distinctive aspects of his pragmatist version of empiricism. The first two of the five points are the basis for Quine’s criticism of Carnap’s doctrine of intensions, as well as a critique of the idea of propositions. The third point, sometimes known as the Duhem-Quine thesis, is the basis for Quine’s critique of logical reductionism, and for his wholistic theses of semantical indeterminacy and ontological relativity. The fourth is his rejection of analyticity, which follows from the third point. And the fifth and final point is Quine’s critique of Carnap’s doctrine of “frameworks” and of the distinction between “internal” and “external” questions. Each of these criticisms is considered in greater detail below.
Quine’s Critique of Intensions and Propositions
At the close of his “Foreword” to Quine’s A System of Logistic Whitehead commented that logic prescribes the “shapes” of metaphysical thought. The logic under consideration of course was that in Whitehead and Russell’s Principia Mathematica, and the metaphysics that is “shaped” by the Russellian syntactical categories – giving the existential claim to the quantifiers – is nominalism. There was probably no expositor of this logic who both illustrated and advocated Whitehead’s comment more consistently than Quine. For more than a decade after System of Logistic Quine published a number of articles which describe how the Russellian symbolic logic and specifically how its theory of quantification enables the user of the logic to exhibit explicitly his ontological commitments, the shape of his metaphysics. The user’s ontological commitment to the kinds of things he believes exists is exhibited by the variable, the symbol that is bound by either the existential or the universal quantifier. The term “variable” in this context has a distinctive meaning that it does not have in mathematics. In his “A Logistical Approach to the Ontological Problem” (1939) reprinted in Ways of Paradox (1966) Quine expresses the rôle of logical quantifiers with the memorable refrain: “To be is to be the value of a variable.” This means that what entities exist from the viewpoint of a given discourse expressed in the predicate calculus of symbolic logic depends on what symbols are bound by quantifiers to become variables.
In 1947 Quine published “On Universals” in Journal of Symbolic Logic and “Logic and the Reification of Universals” in his From A Logical Point of View (1953). In these papers he describes how the nominalist and realist views toward the historic problem of universals are expressed in the Russellian predicate calculus notation. Russell’s convention calls only the particular quantifier the “existential quantifier”, which is indicated by the backward letter “E” symbolically denoted “$“, and expressed in English as “There exists….”. But Quine makes both particular and universal quantifiers express ontological commitments.
The Russellian notational convention is an Orwellian-like nominalist “newspeak”, as it were. The nominalist view is that only individuals exist, and it is expressed in the Russellian notation by limiting the quantifiers to ranging only over the instantiation symbols referencing individual entities. On the other hand the universalist view affirms that attributes or properties exist. In the Russellian notation the existence of attributes is expressed by binding predicates with quantifiers. For this reason Quine calls the universalist view the “platonist” view, and he calls the attributes “abstract entities”. Or when the abstract entities are said to exist in the human mind as meanings or concepts, Quine calls them “mental entities”. A neurologist would likely find this locution appalling.
The Russellian logic thus imposes a distorting dichotomy that reduces both realism and conceptualism to fatuous caricatures that philosophers critical of Plato have dismissed. The notational rôle of the quantifier is referential, such that whatever type of symbol may assume the rôle of a variable bound by a quantifier, thereby assumes the rôle of referencing an entity. Ostensibly Quine’s purpose is not to advocate one or the other ontological thesis, but to advocate the rôle of the quantifiers as making a philosopher’s ontological commitment explicit.
But Quine is not neutral; he has his own emphatic view on the issue of universals. In 1947 he co-authored with Nelson Goodman “Steps Toward A Constructive Nominalism” in The Journal of Symbolic Logic. Unlike most papers appearing in academic journals, this article was not so much an expository analysis, as it was a kind of manifesto advocacy for a nominalist programme for applying the symbolic logic. Quine later denied that he is a nominalist, because he accepts the existence of classes, which he views as a kind of abstract entity. And he accepts the existence of classes, because he could not eliminate them in the “logistic” agenda of reducing mathematics to logic. But he denies that descriptive predicates have any signification with a foundation in reality, and offers no explanation as to why classes are anything but arbitrary collections. Typically nominalists did not reject classes. What they rejected is that there are either mental concepts or real attributes that are the basis for classes, and they view classes as merely collections of entities that are referenced by terms. Thus notwithstanding Quine’s later attempt to separate his views from nominalism, he is a de facto nominalist, because he explicitly rejects the existence of properties, attributes, concepts, ideas and intensions, such as are propounded not only by Carnap but also by the majority of pragmatist philosophers today.
Today philosophers of science investigating scientific revolutions and also those developing computational systems have come to accept the existence of a three-level cognitive semantics of words, both intensions and extensions, concepts and referenced objects, instead of a two-level referential semantics of only words and objects. Nominalists are always troubled by coreferential terms having the same extension but having different meanings or intensions. One reason that Quine rejects these latter types of abstract entities is that they can be eliminated from the logistic reductionist programme as he construes it. The second reason is that he denies that Carnap’s intensions can be treated extensionally, as Carnap attempts to treat them by relating them to classes using analytical statements, a type of statement that Quine rejects.
In “Five Milestones” Quine notes that the first of the five milestones or turning points in the history of empiricism, is the shift from ideas to words. In his Word and Object he calls this shift the “semantic assent”, which he advocates, because philosophical discourse is carried into a domain where participants have greater agreement on the objects, i.e., on words. In “Five Milestones” he says that the shift originated with the medieval nominalists. In fact medieval nominalists such as William Ockham (1285-1347) accepted a three-level semantics including words, signifying concepts and signified forms; what they rejected was the substantial and accidental forms that the earlier Aristotelians and Scholastics said are signified by concepts.
Quine argues against the reification of universals, and says that affirming the existence of abstract or mental entities is due to a common confusion, in which descriptive predicates are given a referential function that is properly had by names and bound instantiation variables. In “Ontological Relativity” (1968) he describes this error as a case of the copy theory of knowledge, which he calls an uncritical semantics. He ridicules this error as the “myth of the museum” and the “fantasy of the gallery of ideas”, by which he means that words are mistakenly understood to be labels for ideas or meanings, as though they were exhibits. He views the confusion between names and descriptions to be a particularly pernicious philosophical error, and he maintains that Russell’s theory of descriptions offers the way to avoid it. This is the technique used by Russell in his “On Denoting” in Mind (1908). In his “On What There Is” (1948) reprinted in Logical Point of View Quine says that Russell’s theory of descriptions enables the philosopher to transform names into predicates, such that names should not be taken as an ontological criterion for deciding what is real. The correct criterion for determining the ontology of a language is the use of the quantified variable, so that predicates are not confused with names, and no claims are made to the effect that predicates name entities, unless the predicates are explicitly quantified.
Closely related to the first milestone, the second is the shift of semantic focus from terms to sentences. In “Five Milestones” Quine explains that the meanings of words are abstractions from the truth conditions of the sentences that contain them, and that it was the recognition of this semantic primacy of sentences that gave us contextual definition. Quine traces the development of contextual definition, which he calls a revolution in semantics, to Jeremy Bentham’s technique of “paraphrasis”, which is a kind of paraphrasing or circumlocution. If Bentham found some terms convenient but “ontologically embarrassing”, contextual definition enabled him in some cases to enjoy the services of the term, while disclaiming its denoting. In “Russell’s Ontological Development” (1966) reprinted in Theories and Things (1981) Quine joins Ramsey’s characterization of Russell’s theory of descriptions as a paradigm of philosophical analysis, and he says that our reward for the paraphrasis technique is the recognition that the unit of communication is the sentence and not the word.
In his Meaning and Necessity Carnap explicitly affirms that intensions are not names either of concepts or of abstract entities. He maintains that like physical properties intensions may be said to be objective without invoking any hypostatization, and that they are indifferent to either concrete or abstract objects. Carnap’s intensions are suggestive of the logicians’ distinction between suppositio and significatio for terms, although Carnap never makes this comparison. According to the theory of supposition a univocal term’s signification or meaning is the same whether the term occurs either as a subject or as a predicate in a categorical proposition. But its supposition as a subject is called “personal”, because it references the individual members of the class according to its associated syncategorematic quantifier, e.g., “all” or “some”, while its supposition as a predicate is called “simple”, because no reference is made to the members of the class it signifies, and its meaning is used indifferently with respect to instantiation. For this reason predicates are not quantified.
Simple supposition enables logicians as well as the ordinary-language user to say, “Every raven is black” and affirm the reality of blackness without also affirming the existence of a Platonic entity called “blackness.” The logician can distinguish names and predicates while still affirming that the descriptive predicates describe something real without hypostatizing it. This capability is denied the user of the Russellian predicate logic, who can only affirm the reality of blackness by quantifying the predicate and therefore treat it as a hypostatized entity. He can only distinguish names and predicates by being nominalist, by denying that descriptive predicates describe anything.
As it happens, when Quine attacks Carnap’s admission of attributes and intensions, as he does in “On the Individuation of Attributes” (1975) in Theories and Things, he attacks Carnap’s use of analytic statements and does not claim that Carnap has confused names and predicates. But even apart from the issue of analyticity, Carnap’s theory of intensions is inconsistent, because he also accepts the Russellian predicate logic with its nominalist notational conventions. In the section of Meaning and Necessity in which he discusses variables, Carnap unfortunately agrees explicitly with Quine’s view that the ontology to which one’s use of language commits oneself comprises simply of the objects that one treats as falling within the range of values of one’s variables, and he explicitly accepts Quine’s refrain that to be is to be the value of a variable. Quine and Whitehead recognized, as Carnap had not, that one’s logic – notational conventions – shapes one’s metaphysics, and Quine’s papers on theory of reference had as their basis the thesis that the Russellian logic affirms existence exclusively by means of the instantiating quantifiers.
The Russellian manner of expressing ontological commitment has its peculiar and controversial aspects, which are clear when contrasted with the earlier Aristotelian logic. In the Aristotelian logic the syncategorematic quantifier does not affirm existence. Instead existence is affirmed by the copula term “is”, as in “Every raven is black.” The noteworthy difference is that in the Russellian notational conventions the only existence that can be affirmed is the entities referenced by the quantified variable, such that any attempt to affirm the reality of attributes or properties must describe them as entities referred to by a quantified predicate. In the Aristotelian logic, however, the reality of what may be called an attribute signified by the predicate need not be hypostatized as some kind of Platonic entity. Quine is therefore consistent with his use of the Russellian logic, when he describes the reality status of red, the property, as an abstract “entity”, and when he describes the reality status of red, the meaning, as a mental “entity.”
According to the syntactical categories admitted by the Russellian logic all philosophers are either nominalists or Platonists, since they must either deny the reality of attributes by not quantifying the predicate, or they must affirm the attributes as Platonic entities by quantifying over the predicate. In the Russellian logic attributes, properties, aspects, and accidents have no reality status except as subsisting entities. Carnap’s attempt to admit intensions or meanings and properties that are not hypostatized is inconsistent with his use of the Russellian logic and with his agreement with Quine that ontology is described by means of bound variables. And his complaint about erroneously labeling philosophers “Platonists” is similarly inconsistent. Other and more consistent philosophers have recognized the Russellian logistic to be an Orwellian-like “newspeak” for advocating a nominalist agenda built into its notational conventions, which the pontificating Quine attempts to enforce by calling it the “canonical notation.”
In his Medieval Logic and Metaphysics (1972) the University of Manchester British philosopher, David P. Henry, asks how modern logic, caught as it is in the “entanglement” of the expression of existence in the quantifiers, can recapture the untrammeled approach to existence enjoyed by its medieval predecessors. He proposes reconsideration of the modern formal logic of the Polish logician Stanisław Leśniewski (1886-1939), who is unfamiliar to most Anglo-American modern logicians. In his autobiography Quine recounts his arguing with Leśniewski about “abstract entities” (Quine’s characterization) when visiting Warsaw in the 1930’s. Henry notes that Leśniewski’s logic employs an interpretation of the quantifiers, which enables their dissociation from its currently conventional entanglement with the notion of existence. Henry gives examples of how Leśniewski’s interpreted system with its ontology may be used in the analysis of medieval themes including supposition with a symbolic logic designed by Henry. In the present context the significance of Henry’s work is that it shows how Quine’s ontological agenda does not imply a simplistic dichotomy between modern mathematically expressed logic and antiquated colloquially expressed Aristotelian logic, but rather depends on the contrived notational conventions distinctive of the Russellian logistic, to which there can and do exist superior alternatives.
Quine’s weltanschauung seen through the lenses of Russellian symbolic notation with its ontological agenda reducing attributes either to “entities” or to unreality is a terminal case of the mathematician’s disease. Quine’s nominalist rejection of properties, attributes and qualities denies such qualitative differentiation its foundation in reality that enables conceptualization. In fact in pursuit of their “logistic” agenda the Russellians firstly had to reduce logic to mathematics before they could reduce mathematics to logic. And it may be added that attempted paraphrasis by quantifying predicates does not evade nominalist ontology; it only incurs a fallacy that Whitehead called “misplaced concreteness”, i.e., the Platonic hypostatization of properties, which earlier logicians had avoided by their recognition of supposition. Also the nominalism built into the Russellian notational conventions by combining existence and quantification is a prior ontological commitment, which is as inconsistent with Quine’s ontological relativity as is his positivist behaviorism.