Relations, Internal and External

views updated


Common sense would seem to hold that if some properties of a thing were taken away from it, it would no longer be the same thing. Further, it seems to hold that this is not the case for all properties of the thing. This intuition is the basis of the distinction between essential and accidental properties of a thing. It is also the basis of the distinction between the internal and the external relations that that thing bears to other things. For if among the properties that are essential to a thing (for example, the state of Maine) are relational properties, properties whose characterization essentially involves reference to some other thing (for example, the property of being north of Boston), then we say that the relations in question (for example, the relation between Maine and Boston) are internal to that thing (Maine). If we think that the thing would be the same were it (for example) not north of Bostonas in the case of a railroad car traveling through Mainethen we say that the relation in question is merely external to that thing.

The most familiar sort of relations considered when the topic of internal relations is discussed are relations between two or more particulars. However, the same internalexternal distinction may be drawn in the case of relations between universals and particulars and also in the case of relations between two or more universals. If one holds that for every property P that a particular X displays, there is a universal, P -hood, to which X stands in the relation of "exemplification," then all of X 's properties may be construed as relational properties. Some of these relations of exemplification may be regarded as internal to X and others as external. Again, one may say that a universal such as "manhood" stands in an internal relation to certain other universals (for example, "rationality") and in an external relation to other universals (for example, "philosopherhood"). Here the internal relation in question will be entailment, in the sense of "entails" in which we say that a given property ("being a man") entails another property ("being rational"). In what follows, however, we shall confine ourselves as far as possible to relations holding between particulars, both because the philosophical literature has focused on such relations and because the notions of "exemplification of universals" and of "relations of entailment holding between universals" are sufficiently obscure and controversial to require detailed supplementary discussions. (Also, we shall not always trouble to distinguish between discussion of internal properties and of internal relations, since whatever doctrine a philosopher holds about the former will apply, mutatis mutandis, to the latter).

Two extreme positions have been put forward by philosophers who regard the internal-external distinction as unclear or incoherent. The first is that all of a thing's properties are essential to its being what it is (and, a fortiori, that all its relations are internal to it). This position is associated with idealism and monism, for reasons that will emerge as we proceed. It holds that the connections between each of a thing's properties (including its relational properties) and all of its other properties are so close that the deprivation of a single property would force us to say that, in a nontrivial sense, the thing is no longer what it was.

The second extreme position holds that none of a thing's properties are essential to it (and thus, a fortiori, that no relations are internal to it). This view is put forward by those who make a firm distinction between the thing itself and a description of it. These philosophers say that, although certain properties of the thing are such that a given description could no longer be correctly applied to it were these properties absent, the notion that "the thing would no longer be the same" if these properties were absent is either trivial or misleading. For, in the weakest sense of "same," the absence of any of its properties would make the thing no longer the same. Any stronger sense will, however, equate "being the same thing" with "being such that a given description correctly applies to it." But since for each thing there are an infinity of equally correct descriptions, and nothing in the thing itself determines which of these is the description, any specification of "essential properties" will be arbitrary.

Both positions hold that the traditional essence-accident distinction, which is drawn by common sense and was first formulated explicitly by Aristotle, must be abandoned. The second position holds that the notion of "essential property" must be seen as a purely conventional notion, without a ground in the nature of the thing itself. It therefore suggests that we replace the notion of a relation being internal to a thing with the notion of a given relational description of a thing (such as "being north of Boston") being internal to (that is, a necessary condition of) another description of the thing (such as "being in Maine"). The first position holds that the notion of "essential property" suggests, wrongly, that there is such a thing as a nonessential property. But since omniscience would see the universe as a seamless web (and, perhaps, as one single individual thingthe Absolute), this suggestion is misleading. Granted, they may say to representatives of the second position, that our present notion of "essential property" is a merely conventional one, we should not be led to conclude that things have no intrinsic natures. They do have intrinsic natures, but these can be known only sub specie aeternitatis, as facets of the Absolute. The commonsense essence-accident distinction is natural and inevitable, given the imperfect state of our knowledge. For omniscience, however, this distinction would be pointless.

This brief sketch of the opposing positions suffices to suggest how intimately the issues about internal relations are bound up with a whole range of other philosophical problemsproblems about the notions of substance, of essence, and of "bare particulars," about "real" versus "nominal" definitions, about nominalism versus realism, about the way in which we refer to and identify particulars, and about the nature of necessary truth. It is perhaps not too much to say that a philosopher's views on internal relations are themselves internally related to all his other philosophical views.

The View that All Relations are Internal

The view that all relations are internal, in the form in which it has been discussed in the twentieth century, originated in the writings of the absolute idealist school in England and America in the period 18901920. In various forms it was held by F. H. Bradley, Josiah Royce, Bernard Bosanquet, and many others. Its most recent sustained defense is found in the work of Brand Blanshard, a follower of Bradley, notably in The Nature of Thought (1939). It has obvious historical connections with the doctrines of the seventeenth-century rationalists, notably Gottfried Wilhelm Leibniz's view that all truths are analytic and Benedict de Spinoza's assimilation of causal relations to logical relations. Its most important historical antecedent, however, is the philosophy of G. W. F. Hegel. Hegel's insistence that the world was rational through and through, because Reason (or "Spirit") alone was real, was the principal inspiration of the philosophers who adopted the view that all relations were internal. For, if some relations were external, then the universe would be "impenetrable" to reason, in the sense that there would be brute particular facts not deducible from universal truths even by God himself.

A. C. Ewing, in Idealism (1934), provides a comprehensive account of the various meanings given to the term internal by exponents of absolute idealism and a critical analysis of those arguments in favor of the doctrine that all relations are internal that depend upon an ambiguous use of "internal." As Ewing points out, the meanings given to "internal" ranged from a very weak sense, in which to say that the relation R which X bore to Y was internal to X meant merely that "R makes a real difference to X," to a very strong sense, in which it meant that "from a knowledge of Y and R we could infer with logical necessity that X possesses a certain determinate or relatively determinate characteristic other than the characteristic of standing in the relation in question." Because such ambiguities permeate the discussion of the topic in such writers as Bradley and Royce, we shall not attempt an exegesis of their arguments. Instead, we shall attempt a reconstruction of two particularly persuasive arguments that seem to represent at least part of the common core of the absolute idealists' defense of their position on this subject. The two arguments to be examined by no means exhaust the repertoire of arguments that have been deployed in favor of the view that all relations are internal, but they are the arguments on which criticism of this view has chiefly centered.

argument from the nature of self-identity

The first argument, which will be called here the argument from the nature of self-identity, was first clearly formulated by a critic rather than a proponent of the view that all relations are internal. G. E. Moore, in a classic attack on this view ("External and Internal Relations"), suggests that "one thing which is always implied by the dogma that 'All relations are internal' is that, in the case of every relational property, it can always be truly asserted of any term A which has that property, that any term which had not had it would necessarily have been different from A." The argument in favor of this view is simply that, as Moore puts it, "if A has P, and x has not, it does follow that x is other than A." In other words, it is unquestionably true that

  1. A has P entails that (x does not have P materially implies that x is other than A ).

Contemplation of this truth, Moore suggested, led philosophers to say that "A could not be what it is (but would necessarily be something different) did it not have P."

Now, as Moore points out, the argument as it stands is fallacious. (1) does not permit the conclusion that

  • (2) A has P materially implies that (x does not have P entails that x is other than A ).

Only (2) would permit the conclusion that A would necessarily be a different particular did it not have P. The difference between (1) and (2) may be put by saying that all that (1) tells us is that A cannot both have and not have the property P, whereas (2) tells us that A could not be A unless it had P. (1) is trivial, whereas (2) blurs the commonsense contrast between essential and accidental properties (and thus between internal and external relations). As Moore puts it, "(1) asserts that if A has P, then any term which has not, must be other than A ; (2) asserts that if A has P, then any term which had not, would necessarily be other than A." Moore notes that to confuse the two propositions, "you have only to confuse 'must' or 'is necessarily' with 'would necessarily be."" This confusion, in turn, will lead one to confuse the (physically necessary but logically contingent) fact that A has P with a statement about what is logically necessary for something to be A. While not attempting to cite examples of this fallacy in the writings of the absolute idealists, Moore claimed that much of their willingness to adopt the view that all relations are internal was due to their having confused (1) and (2). Whether or not this fallacy played the role in their thought that Moore thought it did is less important, from a historical point of view, than the influence exercised by Moore's diagnosis. Philosophers in general tended to agree with Moore that the absolute idealists had been guilty of this confusion, and his essay was a turning point in discussion of the topic. Defenders of the thesis of the internality of all relations who came after Moore were forced to produce arguments against the main presupposition of Moore's argumentthat the commonsense distinction between logically contingent propositions and logically necessary propositions was unobjectionable. Crudely put, one may say that before Moore's essay, defenders of the view that all relations were internal felt able to argue that simple reflection on commonsense criteria for self-identity led to the conclusion they desired. After Moore's essay, they were forced to attempt to undermine common sense by claiming that the distinctions Moore had drawn were, though commonsensical, philosophically indefensible.

argument from the nature of causality

The above was the strategy adopted by Blanshard in his The Nature of Thought, in which he presents the second, far more important and profound, argument in favor of the doctrine that all relations are internal. This may be called the argument from the nature of causality. Moore, like most philosophers in the tradition of British empiricism, had taken for granted a distinction between physical necessity and logical necessity, a distinction between the sense in which it is necessary, given the laws of nature and the past history of the universe, that a given particle be located at a given point in space at a given time, and the sense in which it is not necessary, simpliciter. Traditional rationalism, on the other hand, had questioned this distinction. Although earlier absolute idealists had also rejected the distinction between two kinds of necessity, they had done so en passant. They had treated it as simply one more consequence of empiricism's uncritical acceptance of a commonsense metaphysics that, they claimed to have shown, was fundamentally incoherent. Blanshard, approaching the matter epistemologically rather than metaphysically, brought forward a battery of arguments designed to show that the acceptance of this distinction was the result of a mistaken Humean analysis of knowledge. By weakening this distinction and claiming that causal necessity (by virtue of which A had P ) could not be separated from logical necessity (by virtue of which A was self-identical), he was able to argue that what Moore had viewed as a simple confusion was at worst a confused formulation of a vitally important insight.

In examining this second argument, it will again be convenient to look to its critics rather than to its defenders. Ernest Nagel, in a critique of Blanshard's The Nature of Thought titled "Sovereign Reason," restates and criticizes Blanshard's views on internal relations in a way that brings out very clearly their connection with Blanshard's treatment of causality. Blanshard, in turn, has replied to Nagel in the later chapters (particularly Ch. 12) of his Reason and Analysis (1963). A summary of the Blanshard-Nagel controversy will serve two purposes. It will trace the most recent line of defense adopted by defenders of the view that all relations are internal, and it will lead us to an understanding of why some philosophers claim that no relations are internal.

Blanshard puts forward, and Nagel quotes as a basis for criticism, the following version of the doctrine that all relations are internal. Despite the ambiguities detected by Ewing, Blanshard holds that "the principal meaning" of this doctrine is clear and formulates it as follows:

(1) that every term, i.e., every possible object of thought, is what it is in virtue of relations to what is other than itself; (2) that its nature is affected thus not by some of its relations only, but in differing degrees by all of them, no matter how external they may seem; (3) that in consequence of (2) and of the further obvious fact that everything is related in some way to everything else, no knowledge will reveal completely the nature of any term until it has exhausted that term's relations to everything else. (Nature of Thought, Vol. II, p. 452)

Nagel notes, and Blanshard would agree, that everything here turns on the notion of the "nature of a term." If the term's nature includes all its properties, then Blanshard is right. Nagel bases his general objections to Blanshard on the claim that this is a perverse use of "nature," since "it is quite clear that just what characters are included in an individual, and just where the boundaries of an individual are drawn, depend on decisions as to the use of language. These decisions, though motivated by considerations of practical utility, are logically arbitrary " (p. 275). Nagel, in other words, is saying that "the nature of X " consists of just those properties of X whose absence would cause us to cease using "X " to refer to X and that the selection of these properties is determined not by empirical study but by convention. The list of such properties is finite, whereas the list of the properties of X is potentially infinite. Nagel thus adopts what has become the standard empiricist view, first clearly formulated by A. J. Ayer in "Internal Relations," that to determine which properties of X are internal to it is merely a matter of determining which propositions about X are analytic and that determining this is simply a matter of consulting linguistic usage. To urge that the nature of a thing includes all its properties would, given this view, be to urge that all propositions about X are analytic. Both Nagel and Ayer treat this conclusion as a reductio ad absurdum.

In examining Blanshard's arguments, Nagel first takes up Blanshard's form of the argument from the nature of self-identity and disposes of it by drawing what is essentially Moore's distinction between the logically contingent fact that A has P and the logically necessary fact that anything that does not have P cannot be identical with A. His defense of this distinction is simply that unless the distinction is drawn, we shall wind up with the view that "the nature of X " is identical with X itself and thus that "the nature of a thing, like the thing itself, would be something that is in principle indefinable and could not therefore be made the basis for bringing into systematic order any of the characters which the thing displays" (p. 276). But from Blanshard's point of view, this reply begs the question, since Blanshard would be quite willing to say that the nature of any given particular is indeed indefinable (by finite minds). For Blanshard the question is merely pushed back to the issue of whether a satisfactory epistemology can be constructed on the basis of the view that all logical necessity has its source in linguistic convention. But this latter issue is just the issue of whether causal relationships (which are agreed on all sides to be matters not of convention but of empirical inquiry) can, in the last analysis, be held to be distinct from logical relationships. If they cannot, then it would seem fair to say that although we must (unfortunately) work with the commonsense distinctions between necessary and contingent truths, essence and accident, physical and logical necessity, and the like, these distinctions are nevertheless mere pragmatic makeshifts (pertaining, in Bradleian terminology, to Appearance rather than to Reality). To invoke them to is to attend not to how things are but merely to how we are forced (by the limitations of our minds and of our everyday language) to talk about them.

Thus the battle between Blanshard and Nagel is truly joined only when Nagel takes up the question whether "logical necessity is involved in causal relations." Blanshard has, as Nagel notes, two principal arguments for the view that it is so involved. The first is that causal relations must be analyzed either in terms of "mere regularity of sequence" or in terms of "entailment." The failure of the regularity view will, in Blanshard's eyes, constitute a proof of the entailment view. But the entailment view is just that "A causes B " is a statement about a logical relation between A and B. Now if (as is not implausible) all true relational propositions about particulars are propositions that are true in virtue of causal relations between the particulars mentioned in these propositions, then it follows that all particulars are connected to all others by logical relations and that every such proposition would be seen (by omniscience) to entail a logical truth about every such particular.

Nagel has two objections to this argument. First, the "regularity" and "entailment" views do not exhaust the available analyses of causality; second, "the entailment view contributes nothing toward advancing the aims of specific inquiries into the causal dependencies of physical nature." The second objection can be dismissed by Blanshard as irrelevant, since he is quite willing to admit, with David Hume, that observation of regular sequence is our only method for determining what causal relations actually hold (except, perhaps, in the case of "direct insight" into certain relations between mental states or events). Blanshard need merely insist that regularity provides evidence of an underlying entailment but that the regularity and the entailment must not be confused. Blanshard offers no reply to Nagel's first objection, but one suspects that he would argue that all proposed via media analyses of causality in fact boil down to one of the two alternatives he has suggested. Even if this point is granted to Blanshard, however, the whole question of the validity of his attack on the regularity theory remains. We must leave the topic with the remark that Blanshard can, in attacking this theory, take full advantage of the embarrassment encountered by Rudolf Carnap, Nelson Goodman, and others in their attempts to construct an inductive logic on the basis of Neo-Humean "regularities." Further, recent work in inductive logic (such as Goodman's Fact, Fiction and Forecast, 1955) and the philosophy of science (the work of Hilary Putnam, Wilfrid Sellars, P. K. Feyerabend, and others) has made it apparent that the distinction between matters of convention and matters of fact is not so clear as Hume and the early positivists believed. This recent work is closely connected with W. V. Quine's skepticism about the analytic-synthetic distinction and related work in the philosophy of language. It is perhaps not too much to say that empiricism is presently in a state of crisis and that the crisis revolves precisely around the validity of the distinctions that empiricists have traditionally invoked against the thesis of the internality of all relations. We must conclude that the question of the validity of Blanshard's first form of the argument from the nature of causality must remain undecided until these issues have been further clarified.

Before leaving the Blanshard-Nagel controversy, however, we must take up the second of Blanshard's arguments in favor of the view that logical necessity is involved in causation. This argument is that philosophical reflection upon the nature of causality leads us to conclude that

to say that a produces x in virtue of being a and yet that, given a, x might not follow, is inconsistent with the laws of identity and contradiction. Of course if a were a cluster of qualities abstracted from their relations, and its modes of causal behaviour were another set conjoined with the former externally, then one could deny the latter and retain the former with perfect consistency. But we have seen that when we say a causes x we do not mean that sort of conjunction; we mean an intrinsic relation, i.e., a relation in which a 's behaviour is the outgrowth or expression of a 's nature. And to assert that a 's behaviour, so conceived, could be different while a was still the same would be to assert that something both did and did not issue from the nature of a. (Nature of Thought, Vol. II, p. 513)

With this argument, as Nagel notes, we are back at the perplexing notion of "the nature of a." Whereas the entailment analysis of the nature of causation can perhaps be stated without using the notion of the "nature of A " (although if it were, it might be difficult for Blanshard to infer the thesis of the internality of all relations from the truth of the entailment view), this present argument about the nature of causality makes essential use of this notion. At this point, therefore, Nagel returns to his general line of attack on Blanshard's formulation of the thesis of the internality of all relations and argues that what Blanshard says here is true only if "the nature of X " is defined as "all the properties of X," a definition that, in Nagel's eyes, is both idiosyncratic and such as to trivialize Blanshard's claim.

The effectiveness of Nagel's reply can be judged only in the light of a general theory about the relation between thought, language, and reality. For, here again, Nagel is taking for granted the view that whether a given property is included within a thing's nature is a question about our language, rather than a question to be settled by further inquiry about the thing itself. Just as judgment of the validity of the first form of Blanshard's argument from the nature of causality must be postponed until certain general philosophical issues have been (at least) clarified, so also judgment of the validity of the second form of this argument must be deferred until questions about the standard empiricist doctrine that all "essences" are "nominal" and that "real essence" is an incoherent notion are settled. For Blanshard can insist that Nagel has begged these latter questions. In Reason and Analysis we find Blanshard arguing that Nagel's view that decisions about what characters are included in an individual are "logically arbitrary" leads to the view that, for example, Socrates's snub-nosedness is as good a candidate for an essential property of Socrates as his philosopherhood. Blanshard thinks this a reductio ad absurdum, but this rebuttal, once again, merely moves the argument one step further back. Nagel's point is not that we arbitrarily select which characteristics of an individual shall count as essential but that the criteria of selection are pragmatic, dictated by our present interests and the modes of classification that we have, in the past, found it convenient to adopt. Nagel would say that a choice about linguistic usage, which is, from a practical point of view, far from arbitrary, is nonetheless logically arbitrary, in the sense that a language with alternative conventions is, though inconvenient, perfectly possible.

Blanshard's basic disagreement with Nagel consists in his view that such pragmatic considerations are not the last word and his insistence that the goal of thought is the discovery of real essences. Such real essences would be discovered by discovering the chains of entailment that connect all the various universals that characterize (and, in Blanshard's metaphysics, constitute) a particular. In Blanshard's view, to say that analytic propositions are true by convention is thoroughly misleading, for such conventions are the results of attempts to discover such entailments. For Blanshard the identification of the nature of X with X itself, and of both with the totality of properties that characterize X, and of all of these with X -as-known-by-an-ideal-knower (one who could grasp the entailments between all of these properties), is not (as it is for Nagel) a series of confusions but is forced upon us by an analysis of what we mean by "knowing X." The validity of Blanshard's second form of the argument from the nature of causality ultimately depends upon the validity of this analysis.


The nature and depth of the issues involved in the controversy between Blanshard and Nagel may be further clarified by calling attention to one more area of disagreement between them. This concerns the nature and knowledge of universals. Blanshard views a particular as a congeries of universals and views the internal relations between particulars as reflecting the internal relations holding between the universals that constitute them. It is almost a cliché of recent analytic philosophy that to have knowledge of a universal is simply to know the meaning of a word; thus, to be acquainted with all the universals that characterize a particular would be merely to know the meanings of all the words correctly applicable to that particular. Such knowledge would obviously fall far short of telling us about the relations in which that particular stands to other particulars. For Blanshard, however, universals have natures that are not known to those who merely know the meanings of the words that signify those universals. To know the nature of a universal "fully and as it really is" would involve knowing its relations to all the universals that are exemplified in all the particulars that exemplify the first universal.

Thus, to know any universal "fully and as it really is" would be possible only for omniscience, just as, and for the same reasons that, knowledge of the real essence of a particular would be possible only for omniscience. Thus, resolution of the controversy about internal relations would require, at a minimum, a decision concerning the adequacy of a nominalistic account of universals. Blanshard views the current antagonism toward idealism (and a fortiori toward the thesis of the internality of all relations) as largely a result of analytic philosophy's "systematic confusion between thought and language," a confusion that leads philosophers such as Ludwig Wittgenstein to hold (1) that the notion of having a concept or being acquainted with a universal prior to the use of language is incoherent, and (2) that the notion of detecting internal relations between universals apart from considerations of linguistic usage is a relic of a radically mistaken analysis of mental events. If these latter tenets are accepted, clearly Blanshard's arguments cannot even get off the ground. Once again, we must conclude that the thesis of the internality of all relations cannot be profitably discussed until one has taken sides on the most fundamental issues in contemporary philosophy.

The View that no Relations are Internal

When we turn to the view that no relations are internal, we turn from a controversy that reflects profound underlying disagreements concerning the analysis of knowledge to a controversy about much narrower issues concerning the analysis of naming and predication. Those who say that no particular is internally related to any other particular insist that the only entities that can be internally related to one another are characteristics of particulars. Following to its logical conclusion Nagel's claim that the assignment of a given description to a given particular is "logically arbitrary," they hold that to say that X would "not be what it is" unless it had P is merely to say that the particular could not be characterized in a given way unless it had this property. But since the particular is sublimely indifferent to how it is characterized, it "is what it is" regardless of whatever properties it may have. To speak of "logically necessary conditions for the self-identity of X " is, at best, to speak elliptically of "logically necessary conditions for correctly describing X as a K," where "K " signifies some kind of thing of which X is a representative, or (more generally) of "logically necessary conditions for correctly describing X as C," where "C " is some general characterization.

The whole notion of "properties (and, a fortiori, relations) such that X would cease to be what it is if they were removed" is thus either incoherent or misleading. For "being what it is" is simply too ambiguous a notion; there are indefinitely many kinds to which X belongs and indefinitely many characterizations that apply to it. "Being what it is" is incoherent if it suggests that one of these kinds or characterizations is intrinsically privileged and misleading if a user of the phrase has already picked out some such kind or characterization, thus making his choice "privileged" by stipulation. To philosophers who deny the internality of any relations, the whole notion of internal properties and relations is an unfortunate vestige of the Aristotelian notion that there are real essences of particulars to be discovered by empirical inquiry. These philosophers heartily agree with the seventeenth-century rationalists, and with Blanshard, that any Aristotelian attempt to divide intrinsically essential and intrinsically accidental properties is foolish. But whereas Blanshard, sticking to the quest for real essences, insists that this point merely shows that the real essence of an object must include all its properties, these philosophers take the point to show the incoherence of the notion of "real essence" and the notion of "internal property."

It may be useful to put the contrast between the roughly Aristotelian commonsense view and the two extreme views in yet another way. If we say that common sense holds that there are both particulars and properties of particulars, then we may say that common sense holds that each particular stands in a necessary relation to some of its properties and in a contingent relation to others. Blanshard dissolves the particular into a congeries of properties, and, because he believes (a ) that properties (qua universals) have intrinsic natures to be discovered by inquiry (other than inquiry into linguistic usage) and (b ) that such inquiry would, in principle, discover relations of entailment between all possible properties of all possible particulars, he holds that a particular stands in a necessary relation to all its properties. Philosophers who deny both doctrines and who assert (c ) that "logical necessity" can only characterize relationships between universals, naturally emerge with the conclusion that the whole notion of logically necessary relations between particulars and their properties must be discarded. To put it picturesquely, Blanshard thinks that the dissolution of the traditional essence-accident distinction leaves us with the particular as a node in a network of internal relations between universals. His opponents think that this dissolution leaves us with "bare" particulars on the one hand (particulars that could logically have any properties) and with a network of entailments between universals on the other (a network that is, however, much "looser" than Blanshard's, since between most universals no relations of entailment exist).

concept of bare particulars

As an illustration of the movement toward leaving particulars bare, we may cite Gilbert Ryle, who says, in his article "Internal Relations," that

for this view [the thesis of the internality of all relations] to be true or false, it would have to be significant to predicate a logically proper name or designation of a logically proper name or designation; and it would have to be significant to assert or deny that this was this ; and the question "is anything this?" would have to mean something. "This" is not a predicate, and a sentence in which it pretends to function as one is meaningless. So there could be no such dispute as to whether this's being this does or does not depend on its being in one or other of its relations. (p. 165)

This line of thought suggests the general conclusion that there are no analytic propositions that ascribe properties to particulars. For example, it is misleading to call "Socrates was a Greek philosopher" analytic, for what this statement expresses is either (1) the contingent fact that certain features (snub-nosedness, being married to Xanthippe, and so on) were compresent with certain others (being Greek, being a philosopher), or (2) the contingent fact that the word Socrates is used to refer to an individual who exhibited certain features.

Even among philosophers who both reject (a ) and (b ) and accept (c ), however, this general conclusion has been a matter of debate. In what follows, we shall consider an attempt to avoid the conclusion that there can be no analytic propositions that ascribe properties to particulars and an attempt to avoid the extreme position that no relations are internal to particulars by providing a "rational reconstruction" of the commonsense view. Such attempts are motivated, at least in part, by philosophical discomfort over the notion of "bare particulars." The nature of this discomfort may be illustrated by considering the question "What, then, are these particulars, apart from the properties we ascribe to them?" If particulars really are "bare," then any answer to this question is bound to be either wrong (if it lists some features that are criteria for particularity) or unhelpful (if it consists in saying simply "Well, particulars are just the kind of thing that properties can be ascribed to"). Although the realistic bent of contemporary analytic philosophy makes philosophers hesitate to accept the Bradleian-Blanshardian view that the whole category of (plural) "particulars" belongs to Appearance rather than to Reality, it nevertheless seems that having only bare particulars would be as bad as having no particulars at all.

internal properties as relative

The most explicit and comprehensive attempt to avoid Ryle's conclusion and still retain most of his premises is found in an article by Timothy Sprigge ("Internal and External Properties"); an examination of Sprigge's treatment of the problem will bring out the underlying issues concerning naming and predication upon whose resolution the present question depends. Sprigge notes that the strength of the Rylean position lies in the fact that

in sentences expressing particular propositions where the subject word is a name, the subject word has no connotation. Therefore no predicate word can have a connotation which is incompatible with the connotation of the subject word. But a subject-predicate sentence could only express a necessary proposition if the connotation of the subject word were incompatible with the connotation of the negation of the predicate word. Of course, this rests upon the questionable view that there may be naming words without connotationand this indeed is basically the point at issue. (p. 204)

One reason why this latter point is disputable is, as Sprigge says, that "it seems that one must identify a thing by some description. Having been thus identified," he continues, "as answering to that description, is it not in effect defined as the thing having those properties, which properties therefore it necessarily has?" (p. 205). In other words, proper names could not be used unless their users could identify their referents, and how could the users do this save by having a description in mind? Must we not say that the notion that the logician's dogma that "proper names do not connote" is true only of such Russellian "logically proper names" as "this" (which cannot be used save in the presence of their referents)? Sprigge replies to this point by granting it but noting also that since the same particular can be identified by an indefinitely wide range of different descriptions, the point is useless if one is trying to defend the notion of internal properties. In the case of a predicate, rough agreement on criteria for its application is required if the term is to play a useful role. But there seems nothing to prevent every speaker of the language from having a different set of procedures for identifying a particular while nevertheless using the same proper name for it. Too many connotations are, so to speak, as bad as no connotation at all for purposes of formulating necessary truths.

If we follow Sprigge here, we need not be troubled by the spectacle of bare particulars. Every particular we refer to will always be dressed in some description or other, so we need not worry about how they look when undressed. But since each particular can be dressed up in so many ways, we are as far as ever from understanding what an "internal property" might be, unless we relativize the notion and say that certain properties are internal to X relative to a person S whose personal criteria for identifying X include the presence of these properties. Relativizing the notion in this way is, in essence, the basis for Sprigge's "reconstruction" of the notion of internal property. As a sample of the sort of intuition upon which the commonsense distinction between internal and external properties is based, he notes that even though we are driven by the Rylean reasoning outlined above to call all subject-predicate statements about particulars synthetic, we find it hard to imagine the falsehood of, for example, "Scott was, at some time in his life, a man." But what is a synthetic proposition if not one whose falsehood can be imagined? Sprigge proposes that we simply face up to the fact that there is a class of propositions that, if we must choose between calling them synthetic or analytic, must be called synthetic, even though they do not have imaginable contradictories. Specifically, they are such that no program of empirical inquiry could be formulated that would lead us to decide between them and their contradictories. The point is most effectively made in the following passage:

To ask whether a thing could have been quite different, from what it is, whether Scott could actually have had all the properties of Handel, is on a different level. The questions we have just been asking are all to some degree requests for further descriptions of Scott. But the present question is not one that calls for any investigation of Scott, and it is difficult to accept that a question which calls for no investigation of Scott, to which nothing about Scott is relevant, is really about Scott. (p. 209)

On the basis of these considerations, Sprigge makes the following proposal:

I suggest that a property is internal to a particular to the extent that no information about that particular is conveyed by one who says that it might have lacked that property. I think that the distinction between internal and external properties is not exact. Let F be any property of a thing a. Then F is an external property of a if something interesting and true may be said of the form "if such and such then not-Fa." Otherwise F is an internal property. But as from different points of view different things are interesting, so from different points of view different properties are internal and external. (p. 210)

The notion of "internal" is thus not only made a matter of degree but also relativized to the interests and purposes of those who are discussing X. Conceivably, everyone might be interested in X for a widely different reason; in this case, it would be quite possible that everyone might identify X by means of a widely different, but equally true, description. Then there would be no agreement on internal properties, and an Aristotelian metaphysics would seem unintelligible to us. As it stands, however, we tend to be interested in things for roughly the same reasons and thus to group the same things into the same natural kinds (for example, to regard Scott as "essentially" a man, rather than as a collection of physical particles occupying a given stretch of space time, or as a colorful patch on the landscape of nineteenth-century Scotland). Given this agreement and given our natural taxonomical instincts (our tendency to turn differences of degree into differences of kind whenever possible in order to facilitate inquiry), we can explain the commonsensical character of the distinctions between essence and accident and between internal and external properties (and, a fortiori, internal and external relations).

As an account of the internal-external distinction that avoids both the arbitrariness of Aristotelianism and the counterintuitive character of absolute idealism, Sprigge's proposal is a happy solution. But, like all such solutions, it is no better and no more permanent than the conceptual framework within which it is constructed. There is, to put it mildly, no consensus among philosophers of language as to when a sentence is "about" a given particular, when two sentences are about the same particular, the proper analysis of the notion of "name," the reducibility of names to descriptions, the assimilation of demonstrative pronouns to proper names, the question of whether proper names can be said to have meanings, the utility of the analytic-synthetic distinction, the equation of "necessary truth" with "analytic truth," and a host of related issues. In the absence of a comprehensive philosophy of language in which these issues are clarified and resolved in a systematic way, Sprigge's proposal must be treated as a useful guideline, rather than as a definitive resolution of the issue concerning internal relations. One can imagine, for example, a revivification of the Aristotelian doctrine of predication, according to which "Socrates is a man" exemplifies a radically different sort of predication from "Socrates is a Greek," such an Aristotelian philosophy of language would, when conjoined with a realistic, anti-instrumentalist philosophy of science, produce a view according to which it would make good sense to say that Socrates's humanity really was internal to him, not simply relative to our interests but absolutely and intrinsically. Such a view would argue that "man" signifies a natural kind and is thus naturally suited to be a predicate "in the category of substance," whereas "Greek" or "atoms located at p at t " is not, and that this is an empirical truth.

There probably would never have been a problem about internal relations were it not for the efforts of speculative metaphysicians, such as Parmenides, Spinoza, and Hegel, to undermine our commonsense conceptual framework. If one rejects such attempts out of hand, one will treat the adoption of monism and of the thesis of the internality of all relations as a reductio ad absurdum of the premises from which these views are derived. Since Moore, the vast majority of Anglo-American philosophers have rejected such attempts and have differed only in their diagnoses of the confusions of falsehoods that engendered metaphysical conclusions. As long as the dogma that logical necessity was a matter of linguistic convention remained unchallenged, a simple and elegant resolution of the problem of internal relations seemed possible. However, recent doubts about this dogma (combined with the realization that Aristotle's distinction between essential and accidental properties is not simply a philosopher's invention but is firmly grounded in common sense) have made the problem look more complex than it appeared in the days of Ayer's Language, Truth and Logic. Philosophers who wish, as P. F. Strawson has put it, to substitute a "descriptive" metaphysics for a "revisionary" one are now faced with the problem of reconciling (a ) the existence of this commonsense distinction with (b ) the standard empiricist view that knowledge of how we speak either does not reveal anything about the nature of the objects we refer to, or at least does so in a very different way than does empirical research directed to those objects themselves, (c ) the fact that the meaning we assign to a term is in part a function of the amount of empirical knowledge we possess, and (d ) the fact that common sense seems to require a realistic, rather than an instrumentalistic, view of what it is to "know the nature of an object."

If the difficulties of such a reconciliation prevent "descriptive" metaphysicians from carrying out their chosen task, then the door will be open once again to the two extreme views examined above. It may turn out that common sense is, if not as incoherent as Parmenides and Bradley thought it, at least sufficiently inconsistent as to require the adoption of paradoxical philosophical theses. Whether one then turns to the extreme represented by Ayer's radical conventionalism and instrumentalism, or to the extreme represented by Blanshard's idealistic monism, will be largely a matter of taste. Both views, as suggested above, are parts of internally consistent philosophical systems. Each system retains certain portions of our commonsense framework and insists on these at the expense of other portions. In the absence of a touchstone other than common sense, it is difficult to see how a rational choice between such systems can be made.

See also Absolute, The; Aristotle; Ayer, Alfred Jules; Blanshard, Brand; Bosanquet, Bernard; Bradley, Francis Herbert; Carnap, Rudolf; Common Sense; Goodman, Nelson; Hegel, Georg Wilhelm Friedrich; Leibniz, Gottfried Wilhelm; Moore, George Edward; Nagel, Ernest; Parmenides of Elea; Putnam, Hilary; Quine, Willard Van Orman; Realism; Royce, Josiah; Ryle, Gilbert; Sellars, Wilfrid; Spinoza, Benedict (Baruch) de; Strawson, Peter Frederick; Universals, A Historical Survey.


Allaire, Edwin B. "Bare Particulars." In Essays in Ontology. Iowa Publications in Philosophy, Vol. I, 1421. The Hague, 1963.

Alston, William P. "Internal Relatedness and Pluralism in Whitehead." Review of Metaphysics 5 (19511952): 535558.

Anscombe, G. E. M. "Aristotle." In Three Philosophers, by G. E. M. Anscombe and P. T. Geach. Ithaca, NY: Cornell University Press, 1961. Gives a sympathetic treatment of Aristotle's distinction between secondary substance and quality.

Ayer, A. J. "Internal Relations." PAS, supp., 14 (1935): 173185. Reprinted in Ayer's Language, Truth and Logic. London: Gollancz, 1936. Ch. 8.

Blanshard, Brand. The Nature of Thought, 2 vols. London: Allen and Unwin, 1939; New York: Macmillan, 1940.

Blanshard, Brand. Reason and Analysis. La Salle, IL: Open Court, and London: Allen and Unwin, 1962

California, University of. Studies in the Problem of Relations. University of California Publications in Philosophy, Vol. XIII. Berkeley, 1928.

Chappell, Vere. "Sameness and Change." Philosophical Review 69 (1960): 351362. On criteria of self-identity.

Church, Ralph W. "On Dr. Ewing's Neglect of Bradley's Theory of Internal Relations." Journal of Philosophy 32 (1935): 264273.

Ewing, A. C. Idealism. London: Methuen, 1934. This contains the best exposition of the similarities and differences between the exponents of absolute idealism and also of the relation between the doctrine that all relations are internal and other idealistic doctrines.

James, William. Essays in Radical Empiricism. New York: Longman, 1912. Ch. 3, "The Thing and Its Relations." On Bradley.

Moore, G. E. "External and Internal Relations." PAS 20 (19191920): 4062. Reprinted in Moore's Philosophical Studies. New York: Harcourt Brace, 1922.

Nagel, Ernest. "Sovereign Reason," in his collection of articles Sovereign Reason. Glencoe, IL: Free Press, 1954.

Rome, Sydney, and Beatrice Rome, eds. Philosophical Interrogations, 219246. New York: Holt, Rinehart and Winston, 1964. Largely a continuation of the debate between Nagel and Blanshard.

Royce, Josiah. The Spirit of Modern Philosophy. Boston: Houghton Mifflin, 1892. Expounds the relation between early rationalism and Hegel, as it was conceived by the absolute idealists.

Russell, Bertrand. "The Monistic Theory of Truth." In his Philosophical Essays, 150169. London: Allen and Unwin, 1910. Criticism of absolute idealism.

Ryle, Gilbert. "Internal Relations." PAS, supp., 14 (1935): 154172.

Sellars, Wilfrid. Science, Perception and Reality. London, 1963. Ch. 9, "Particulars." On criteria for self-identity of particulars.

Sprigge, Timothy, "Internal and External Properties." Mind 71 (1962): 197212.

Stace, W. T. The Philosophy of Hegel. London: Macmillan, 1924. Reissued New York: Dover, 1955. Ch. 2 discusses the relation between early rationalism and Hegel, as it was conceived by the absolute idealists.

Thompson, Manley. "On the Distinction between Thing and Property." In The Return to Reason, edited by John Wild. Chicago: Regnery, 1953. Another defense of Aristotle's distinction between "secondary substance" and "quality."

Will, Frederick. "Internal Relations and the Principle of Identity." Philosophical Review 44 (1940): 497514.

Wollheim, Richard. F. H. Bradley. Harmondsworth, U.K.: Penguin, 1959. Ch. 3 contains an exegesis of Bradley's treatment of relations.

Richard M. Rorty (1967)

About this article

Relations, Internal and External

Updated About content Print Article