| Andrew Chrucky, Critique of Wilfrid Sellars' Materialism, 1990 |
CHAPTER 3
ONTOLOGICAL COMMITMENT
For Sellars the task of ontology is to say what sorts of entities suffice for a complete description and explanation of the goings on in the universe relative to some linguistic framework, and ultimately relative to our present linguistic framework. This is so because descriptions and explanations are given in some language or other, and the task becomes one of determining which language suffices for the description and explanation of the panorama of the universe. Once this linguistic turn is taken, the attempt must be made to find a linguistic medium sufficient to express all one wants to.
Some have taken ordinary language as a guide to ontology, while others have tried to remedy, what they took to be, the faults of ordinary language. Sellars is somewhere in between when he introduces the notion of the Manifest Image (which as I have stressed is not identical to a Common Sense framework). But inasmuch as the Manifest Image is intended to capture the major features of common sense, Sellars is both a codifier and rectifier of ordinary language. He believes that for the pragmatic considerations of using ordinary language, one must be sensitive to the "forms of life" in which the language is used; but for ontological and semantical considerations a canonical language is to be sought. This dual consideration is evident is Sellars' debate with Strawson on the problem of presuppositions of definite descriptions.{1} Sellars distinguishes the dialectical (pragmatic) uses of language in which what Strawson has to say about truth-value gaps is correct, and the ontological and semantical uses, for which Russell's theory of descriptions is more appropriate. His final pronouncement is that the Manifest Image is pragmatically useful, but ontologically inadequate.
The problem of ontological commitment is relevant to the regimentation of the Manifest Image and, consequently, to the broader issue of how to combine the Manifest and the Scientific Images. More directly, a knowledge of the ontology of the Manifest Image is important in knowing which resources are available to us in constructing a theory of perception, and, specifically, in determining whether an adverbial theory of perception is appropriate at the Manifest level. The problem of ontological commitment is linked with the notion of ontological reduction. And a key thesis of Sellars' ontology is that in the Manifest Image we are committed to the existence of physical objects, but not events; and this is so because talk of events is reducible to talk about physical objects. To understand why a reduction of events is needed in Sellars approach, we have to focus on what, I take it, is a universally agreed on assumption -- at least by the followers of Bertrand Russell -- that unless a qualification is made, or the context suggests the qualification, or language is used figuratively, a literal subject-predicate (including a relational) statement implies that the subject term refers to an object, or, put otherwise, makes an ontological commitment.
If we seriously take the criterion of ontological commitment (of a perspicuous or literal language) as revealed by the grammatical subject term, then the way to escape from an undesirable ontological commitment is to recast the form of the sentence in such a way that the problematic subject term is no longer the subject of the sentence. The recast version can then be called the 'logical form' of the original 'misleading form'. Sellars' major concern is to logically reconstruct sentences to perspicuously display their logical form.
The distinction between grammatical and logical form was introduced to escape from being committed to the existence of unwanted entities. This technique of recasting sentences into a logical form was, as far as I am aware, first programmatically used by William Ockham.{2} However the first self-consciously systematic use of this technique was by Jeremy Bentham, who called it "paraphrasis."{3} This was, I believe, the same technique used by Russell in his theory of descriptions.{4} And it was and is a standard technique used, I would guess, by most analytic philosophers. What goes by the name 'reduction', in one sense of that term, is just this tactic of paraphrase. The goal of eliminating reference to abstract entities by this technique was practiced also by G. E. Moore, G. Ryle, R. Carnap, T. Kotarbinski, and, more recently, by W.V.O. Quine, D. Davidson, R.M. Martin, and W. Sellars.
Since Sellars is in many areas a follower of Russell, he adopts this principle that the apparent minimal criterion of ontological commitment of a language resides in what is expressed by the grammatical subject of a sentence. "Anything that can be talked about is an object."{5} If the grammatical subject-predicate form is taken seriously as committing us to a referent of the subject term, then the unqualified logical outcome of this kind of thinking would be Meinong's ontology, which gives to every purported object of reference a mode of being or existence.
Now whether Meinong's ontology entails a contradiction, as Russell contended, is disputable; but that it seems to offend our sense of reality, as Russell also contended, is more plausible. To avoid linguistic commitment to unwanted entities, the procedure is to use paraphrase (translation) so that the unwanted entity is no longer mentioned by the subject of the sentence; and yet the transformed sentence is equivalent in some sense{6} to the original one.
But Sellars is quick to recognize that ontological questions cannot be decided simply by paraphrase. After all paraphrase is a two-way street. For example, Sellars believes that in the thing-framework it is physical things which primarily exist, while in the event-framework it is events, i.e., events are reducible to things in the former framework, while things are reducible to events in the latter framework. Whether primacy is given to substances or events must be decided on grounds other than availability of paraphrases. So extra-semantical issues must be considered which guide paraphrase.
Let us generalize and call all efforts to find ontological primacy 'ontological regimentation.' And I will distinguish semantical regimentation from extra-semantical regimentation.
Paraphrase and the search for perspicuous languages loom large in Sellars' philosophical arsenal in revealing ontological commitment. And when talk turns to ontological commitment, one has to deal with Quine. Sellars disagrees with Quine about the criterion of ontological commitment relative to the use of quantifiers in symbolic logic. For him, unlike for Quine, variables do not have a referring role. Variables do not function as indexical expressions, they function as place holders for linguistic expressions. This substitutional approach to quantification theory, carrying no ontological commitment, allows him to quantify over names, predicates, relational predicates, and sentences as well. Ontological commitment for Sellars lies in names only.{7}
I. SEMANTICAL REGIMENTATION
These semantical regimentation efforts take two forms, depending on the apparent entity being dealt with. These are the first-order and second-order reductions.
In the first method, which may be called first-order (or same level) reduction, the troublesome statement is substituted with an equivalent statement in which the troublesome term no longer appears. Sometimes this is described as providing a contextual definition. Russell's theory of descriptions, for example, is in terms of contextual definitions.
To illustrate how paraphrase is used to formulate ontological commitment to physical objects and persons, Sellars offers as examples three types of first-order reductions.{8} The first example is this. "The average man is 5'10" can be paraphrased as "The average number of the heights of n men is the sum of the heights divided by n." The commitment here is to numbers rather than to the average man, and unless numbers are to be included as basic entities, they too must be further reduced. The second example given is: "The elephant . . ." is reducible to "all elephants . . ." The third example is this. Sentences containing the mention of conjunctive individuals as in "Jack and Jill and Tommy are a family" are reducible to "Jack is married to Jill. Jack is a man. Jill gave birth to Tommy." In the second method, which may be called second-order reduction, it is claimed that a semantical statement (i.e., a meta-statement about an object language) is masquerading as a first-order object statement. It was Carnap's contribution to make much of this second technique, which is called by Quine 'semantic ascent'. Sellars uses this technique primarily in service for his reistic (Materialistic) nominalism. Judging by his practice, his endeavor is to show that there need be no logically reconstructed (regimented) sentences such that their subjects refer to abstract entities; that all such sentences can be translated into sentences having subjects which refer to concrete entities. From this perspective, the problem of universals (specifically the realist-nominalist controversy) can be viewed as the question whether all apparent referents to abstract entities are paraphraseable to referents about linguistic entities. The realists say no; the nominalists keep trying.
Sellars uses paraphrase and semantical ascent as means to disengage us from a belief in abstract entities in general, and events in particular. The resulting ontology of the Manifest Image is reistic and nominalistic.
II. EXTRA-SEMANTICAL REGIMENTATION
There seem to be three extra-semantical standards used by Sellars for ontological commitment.
The first extra-semantical consideration with which Sellars expresses sympathy is found in Strawson's Individuals.{9} In Strawson's hands, the question of primacy takes the form of the question: Which entities are independently identifiable? The working assumption in Strawson's work is that whatever is identifiable is also publicly learned. There is a remarkable coincidence between Strawson and Aristotle on which entities have pride of place. Such entities are said to be substances, and are found to be physical objects and persons. Pride of place to questions of ontology in the Manifest Image for Sellars must be given to the Aristotelian-Strawsonian search for primary beings.
Although Sellars uses this criterion, he also introduces another criterion for demarcating the Manifest Image. The Manifest Image is to play the role of one pole as a heuristic device for understanding the history of philosophy. By this criterion, the regimentation of common sense must reflect the practice of, to use Strawson's phrase, 'descriptive metaphysics'.
But when it comes to a criticism of the Manifest Image and its replacement by the Scientific one, Sellars adds the criterion of the best explanation. Evidently, in the Manifest Image those entities which are descriptive of our world, and by which we locate ourselves in the world, are also used to explain the goings on of our world on the Manifest Image level. However, once the Scientific Image is introduced, this coincidence drops out. The criterion of what exists becomes whatever it is necessary to posit in giving the best possible explanation of phenomena. And for Sellars this is to be determined by theoretical science.
[Go to Chapter 4]