The Importance of ‘Being Earnest’

Reference to properties is normally achieved by the use of nominalizations of predicative expressions. The article examines the relation between different kinds of such nominalizations: while traditionally, the terms ‘wisdom’ and ‘the property of being wise’ were thought to be coreferential, there are certain contexts in which they do not seem to be interchangeable salva veritate. Based on this observation, Friederike Moltmann recently developed a theory according to which abstract nouns such as ‘wisdom’ do not refer to properties. It is argued that her theory is flawed and that the existence of the problematic contexts should be explained in non-referential terms. This paper examines aspects of the ontology of properties employed in ordinary discourse. For the purpose of talking about properties, natural languages allow to generate designators of properties by the means of nominalization. In section 1, I shall give a survey about different sorts of such nominalized property designators and formulate the Uniformity Thesis, according to which the different sorts of nominalization are merely linguistic variants to talk about the same things. While the Uniformity Thesis had been the standard view for quite a time, Friederike Moltmann recently attacked it on the basis of some intriguing linguistic data. After introducing her relevant claims in section 2, I argue in section 3 that they are mistaken, and I offer an alternative explanation of Moltmann’s data in section 4. A note in advance: as to avoid the danger of contaminating my own linguistic data with any of my prevalent philosophical (mis)conceptions, I collected examples from the prose of those who mastered the English language far better than myself. These are not data invented to prove a philosophical point, but data about how non-philosophers actually talk, when they talk about properties. 1 For most of the literary quotations I rely on a database of English prose from the fourteenth to the early twentieth century (The Digital Library of English and American Literature, Frankfurt a.M.: Zweitausendseins, 2004). The Importance of ‘Being Earnest’ Page 2 / 17 1. Nominalizations as Property Designators a. A Survey of Property Designators Philosophers most often use one of two kinds of expression when they intend to refer to properties: abstract nouns, such as ‘wisdom’, and complex phrases of the form ‘the property of being F’. But it is instructive to realize that there is a bigger variety of expressions that are apt for identifying reference to properties. In general, if we can combine a term t with the predicate ‘is a property’ (or: ‘is an attribute’, ‘is a quality’, etc.) such as to produce a true statement, then t has at least one use in which it designates a property. Thus, the truth of the following statements leads to the recognition of three different kinds of property designators, namely abstract nouns, gerunds, and ‘to’-plus-infinitive phrases: (1) Wisdom is a property. (2) Being wise is a property. (3) To be wise is a property. We can discover two further kinds of property designators by realising that we may combine the grammatical subjects of (1) and (2) with either a formal prefix (e.g. ‘the quality of’ or ‘the property of’) or an appropriate material prefix (e.g. ‘the virtue of’), such that the resulting expressions qualify as property designators themselves: (4) The virtue of wisdom is a property. (5) The quality of being wise is a property. What is often overseen is that we can also use a that-clause to designate a property, as in: (6) He hath this property of an honest man that his hand is as good as his word. (Thomas Fuller: The holy state / The profane state) Last but not least, we can use ordinary definite descriptions to denote properties, as in: (7) The property Socrates was most famous for is a property (to wit, it is wisdom). 2 As e.g. Strawson once pointed out; see p. 129 of his Subject and Predicate in Logic and Grammar (London: Methuen & Co, 1974). 3 Another example of this kind would be: It is one of the properties of a triangle that the three interior angles taken together are equal to two right angles. (J. Norman Lockyer: Elementary Lessons in Astronomy) Both quotations can be found under the entry for ‘property’ in the Oxford English Dictionary (2 edition, Oxford: Clarendon Press, 1989). The Importance of ‘Being Earnest’ Page 3 / 17 Let me add two remarks on the introduced test and its results: Firstly, I should point out that although terms of the mentioned classes can be used as property designators, they are not always so used. Most of them exhibit also other uses. This is particularly obvious in the case of gerunds, that-clauses and ‘to’-plusinfinitives: whatever role they play in the following sentences Being the morose kind of guy, he used to live on his own, He wanted her to be a saint, He always thought that he would never marry, it is certainly not that of a property-designator. But the same is also true for at least some abstract nouns: ‘wisdom’ and ‘beauty’ are not used as property designators in the sentences ‘That modesty is rewarding is an old wisdom’ and ‘A beauty appeared at my window’; instead, they are used as general terms applying to wise thoughts and beautiful people respectively. When I will nevertheless speak as if certain terms (instead of: terms in certain uses) could be classified as property designators, it is only for the sake of brevity. Secondly, I admit that some of the test statements I presented, namely those in which the predicate is the bare ‘is a property’, sound somewhat awkward and shallow; an example in question is sentence (3). But we should not conclude that such sentences are semantically defective and that the contained nominalizations are not property designators after all. Rather, the awkwardness of such sentences can largely be explained by the fact that they are highly uninformative. The reason is that in many cases, the knowledge that a given term (in a certain use) denotes a property seems to be a condition of being a competent user of that term (in that use). This feature of a term (which we may call its categorial transparency) is in particular exhibited by the kinds of property designator used in statements (1) to (6). Ordinary definite descriptions of properties, however, may or may not possess it; while some descriptions – see (7) – incorporate information about the ontological category of the entity that is denoted, some others stay silent on this issue, as the description used in: (8) What I am thinking about at the moment is a property. The categorial transparency of property designators contributes to the awkwardness of the statements (1)–(5) and (7), as the contrast between them and the informative (6) and (8) shows. Anyway, we can easily generate less shallow variants of the test sentences (1)–(5) and (7): we can enrich the predicates by some modifier and use the more informative result, such as ‘is an interesting/nice/useful etc. property’ or ‘is a The Importance of ‘Being Earnest’ Page 4 / 17 property shared by all philosophers’. In fact, many smooth examples of such predications can be found, as for instance the following ones: Nay; children, to be null is merely a negative property [...] (James F. Cooper: The Prairie) It is the quality of reason to be invariable and constant [...] (James Boswell: Life of Johnson) But the correct applicability of predicates such as ‘is an interesting/nice/negative etc. property’ to some subject term t obviously justifies the classification of t as a property designator. So, the above test works in spite of a certain awkwardness of some cases (which could be avoided by modifying the test a little bit, although I see no need to actually do that). b. The Uniformity Thesis The property designators used in (1)–(3) and (6) are derived terms that are generated by some form of nominalization from either (i) a general term (the designator in (1) derives from ‘wise’), or (ii) a predicate (the designators in (2) and (3) derive from the predicate ‘is wise’), or (iii) a whole sentence (the that-clause in (6) derives from the sentence ‘his hand is as good as his word’). Furthermore, the property designators in (4) and (5) are also nominalizations, but enriched with an additional prefix that adds some classification of the referents of the terms. Such prefixes are best understood as a kind of appositive constructions comparable to ‘the number’ in ‘the number 7’ or ‘the poet’ in ‘the poet Burns’. Although the pure nominalizations ‘wisdom’, ‘being wise’ or ‘to be wise’, and ‘that she is wise’ are naturally construed as being derived from terms of different syntactic categories, these terms have nevertheless something important in common: the expressions ‘wise’, ‘is wise’, and ‘she is wise’ contain the same descriptive material and differ only in their syntactic category, which is due to the presence of the copula and/or the pronoun. It seems plausible to suppose that designators being thus 4 On this view cp. N. Wolterstorff, On Universals (Univ. of Chicago Press, 1970), pp. 70ff.; J. Levinson, ‘Properties and Related Entities’, Philosophy and Phenomenological Research, 39 (1978), pp. 1–22, at pp. 9f.; D. Wiggins, ‘The Sense and Reference of Predicates: A Running Repair to Frege’s Doctrine and a Plea for the Copula’, Philosophical Quarterly, 34 (1984), pp. 311–328, at p. 320; B. Schnieder, ‘Canonical Property Designators’, forthcoming in American Philosophical Quarterly. The Importance of ‘Being Earnest’ Page 5 / 17 derived from the same descriptive material are just different devices for talking about the same things. Let us call this supposition the (Uniformity Thesis) Any two property designators that are derived from the same descriptive material (in the same use), refer to the same object. It should be noted that the Uniformity Thesis asserts only sameness of reference of two property designators with the same origin, but not sameness of sense. The latter is certainly not given between pure nominalizations and those with a material prefix (‘wisdom’ versus ‘the virtue of wisdom’) because the prefix adds its sense to the sense of the complex phrase, albeit not affecting its reference – unless, I hasten to add, the prefix is ill-chosen for the prefixed term, which would render the whole phrase empty (‘the vice of wisdom’ arguably lacks a referent and therefore it is not coreferential with ‘wisdom’, nor with ‘the virtue of wisdom’); I henceforth concentrate on more ordinary cases where the prefix does apply to the referent of the prefixed phrase. Having realized that a material prefix affects the sense of the whole phrase, it may – for an analogous reason – also be doubted whether a pure nominalization and a nominalization with a formal prefix have the same sense: the sense of ‘property’ obviously enters into the sense of ‘the property of wisdom’; but it is at least debatable whether it also enters into the sense of ‘wisdom’ itself. 2. Moltmann’s Charge Against the Uniformity Thesis Although the Uniformity Thesis is attractive and widely held, it has recently been attacked by Friederike Moltmann in a series of papers. Three pivotal theses of Moltmann are (I will not discuss every aspect of her overall account, but concentrate on those of her theses which I take to be mistaken): (M1) Abstract nouns (e.g. ‘wisdom’) do not refer to properties. (M2) Abstract nouns refer to kinds of tropes. (M3) Kinds of tropes are not properties. 5 Cp. vol. II, §136, p. 51, of B. Bolzano, Wissenschaftslehre (Sulzbach: Seidel, 1837; repr. as vols I.11–I.14 of: Gesamtausgabe, Stuttgart: Frommann-Holzboog, 1985–...). 6 While she already argued against the Uniformity Thesis in her ‘Nominalizing Quantifiers’, Journal of Philosophical Logic, 32 (2003), pp. 445–481, at pp. 457–461, she did so at greater length in ‘Properties and Kinds of Tropes: New Linguistic Facts and Old Philosophical Insights’, Mind, 113 (2004), pp. 1–41. Cp. furthermore her ‘Two Kinds of Universals and Two Kinds of Collections’, Linguistics and Philosophy, 27 (2005), pp. 739–776. The Importance of ‘Being Earnest’ Page 6 / 17 Moltmann argues for this controversial triad of claims on the basis of some intriguing linguistic observations concerning the use of a variety of nominalizations. In the remainder of this paper, I will make a case for the denial of thesis (M1) and defend the Uniformity Thesis against Moltmann’s charge. With my denial of (M1), I do not intend to repudiate (M2) as well; this thesis can be held independently of (M1), by denying (M3) and regarding properties as kinds of tropes. I will henceforth concentrate on a discussion of (M1), leaving aside questions about tropes. Before I challenge Moltmann’s account, let me give a short description of how she arrives at her position: Moltmann points out that there is a contrast between many statements which differ only in that one of them has an abstract noun as its grammatical subject, while the other has the corresponding gerund plus categorial prefix as its subject. Some examples of the kind with which Moltmann illustrates this claim are (they are but three of a whole number of similar cases that Moltmann came up with): (9) Shyness is nice. (9*) The property of being shy is nice. (10) Earnestness is important. (10*) The property of being earnest is important. (11) Wisdom is rare. (11*) The property of being wise is rare. Sentence (9) amounts, roughly, to the claim that shy behaviour or shy attitudes are nice (or perhaps that shy people are nice, as far as their shyness is considered). But (9*), Moltmann holds, is not an appropriate vehicle for the same claim. It might, at best, be understood as a claim that probably only philosophers would make; one could, for example, think that the said property is nice because of its formal structure, or perhaps because of certain relations that it bears to other abstract objects. Similarly 7 Because of similar linguistic data, Levinson defended theses akin to (M1)–(M3): abstract nouns, he argued, refer to qualities (e.g. wisdom), conceived of as abstract stuff that can be portioned into particularised qualities. Properties such as being wise, on the other hand, have the character of conditions that can only be fulfilled or not. See his ‘Properties and Related Entities’, p. 10, and ‘The Particularisation of Attributes’, The Australasian Journal of Philosophy, 58 (1980), pp. 102–115, at pp. 106ff. 8 Philosophers who endorse both (M2) and the Uniformity Thesis include Strawson and Wolterstorff; for relevant passages see P.F. Strawson, Individuals (London, New York: Routledge, 1959), p. 168, Subject and Predicate, pp. 129–132, and Wolterstorff’s On Universals, pp. 78f. and 130–139. 9 See ‘Properties and Kinds of Tropes’, pp. 7f., and cp. ‘Two Kinds of Universals’, pp. 745ff. The Importance of ‘Being Earnest’ Page 7 / 17 for (10) and (10*): while the first of them is naturally understood as saying, roughly, that earnest behaviour is important, the latter of them presumably makes sense only as the tenet of some philosopher. Perhaps, someone working on ontology might utter it, because he uses the property of being earnest as his standard example of a property (and cannot think of using a different one). Things are even worse with the third case: sentence (11*), Moltmann holds, could not even be used to express some strange philosophical claim, because instance-distributing predicates such as ‘is rare’ are not acceptable in combination with explicit property-referring terms. Now, Moltmann thinks that the differences between sentences of the cited pairs are best explained as differences in what entities the statements are about. Their subject terms refer to different entities, and indeed, to entities about a different sort. Since expressions such as ‘the property of being shy’ evidently refer to properties, abstract nouns do not. They rather refer to kinds whose instances are denoted by complex phrases such as ‘Jean’s courage’ or ‘Oscar’s shyness’ – things which have been called individual moments, tropes, or modes (to name just some of a broad variety of technical terms employed for such entities). This is a sketch of the reasoning behind Moltmann’s theses (M1) to (M3). Before discussing these central tenets of Moltmann’s account, let me briefly remark on a further but related aspect of it that rather baffles me. If somebody takes the difference between (9) and (9*) to be indicative of a difference in reference (of their subject terms), it is just fair to demand of her an account of what the different referents are and what distinguishes them. Moltmann’s account of this difference is expressed in the following thesis: (M4) Referents of abstract nouns are kinds of tropes. Such entities are not objects, i.e. they do not have properties. But this strikes me as rather mysterious. It seems to me a plain truism that whatever there is, must have some properties. Moltmann would apparently deny this; but her position raises many questions to which there seem to be no answers: why are some entities unable to exemplify properties? what is it about their nature that deprives them of this rather basic capacity? What is worse, Moltmann’s position seems to be self-undermining: she accepts that expressions of the form ‘the property of being F’ refer to properties (which is only reasonable because she is concerned with the very abundant ontology of properties manifested in ordinary language, not with sparse philosophical conceptions such as Armstrong’s). Now, according to Moltmann, non-objects are unable to have 10 ‘Properties and Kinds of Tropes’, p. 21. 11 Cp. ‘Properties and Kinds of Tropes’, pp. 1f., 24, et passim. The Importance of ‘Being Earnest’ Page 8 / 17 properties. But then, they should have at least the following property: that of being unable to have properties. But so, every non-object would have to both lack properties and yet possess one. This seems to be a reductio of Moltmann’s thesis. I may add that Moltmann wavers between different formulations of thesis (M4). She sometimes uses a formulation that sounds weaker than (M4), when she says that non-objects must inherit their properties from other entities; this claim seems incompatible with (M4), because if something inherits a property P, it certainly has P. Anyway, a variant of my criticism applies also to the weaker claim; if an entity x only possesses properties that it inherits from other entities, this should also be true of its property of possessing only inherited properties. But this property is obviously not inherited itself. Be this as it may, I will, in what follows, argue for the Uniformity Thesis, both by direct reasoning and by undermining Moltmann’s arguments to its denial. Thus, I will make a case against (M1); since (M4) only assists (M1), I can avoid further discussion of (M4). 3. Against Moltmann’s Position a. An Argument and Moltmann’s Retreat In order to explain the differences between (9) and (9*) and the like, Moltmann assumes that abstract nouns (such as ‘shyness’) and abstract nouns cum categorial prefix (such as ‘the property of being shy’) refer to different entities. But there is a straightforward argument against her thesis: we can talk not only of shyness, but also of the property of shyness. Since the following claim is true by any ordinary standards: (12) Shyness is a property, the following identity statement is hardly deniable: (=1) Shyness = the property of shyness. But since ‘the property of shyness’ and ‘the property of being shy’ have exactly the same uses and do not give rise to any such differences as those obtaining between (9) and (9*), we should accept that: (=2) The property of shyness = the property of being shy. By transitivity of identity, we immediately arrive at: 12 ‘Properties and Kinds of Tropes’, p. 2. 13 Cp. Wolterstorff, On Universals, pp. 78f. The Importance of ‘Being Earnest’ Page 9 / 17 (=3) Shyness = the property of being shy. However, in her proposal Moltmann provides the material for a response to this argument. She realises that we can build true statements by combining abstract nouns with predicates like ‘is a property’. While this seems to show that abstract nouns denote properties, Moltmann rather concludes that ‘is a property’ is a reference shifting predicate: as employed in (9) Shyness is nice, the term ‘shyness’ refers to shyness. But it does not do so as employed in (12) Shyness is a property. Here, it refers to the property of shyness. So, Moltmann supplements her theses with the following claim: (M5) In combination with abstract nouns, the predicate ‘is a property’ generates a shift of reference (such that abstract nouns then refer to properties, contrary to what they do outside the scope of such a predicate). This treatment, ingenious as it is, strikes me as rather artificial; what is worse, it suffers from some lethal deficiencies which I shall now propound. b. Against the Shift-of-Reference Thesis If a certain phrase p generates a reference-shifting context for some term t (a context in which t has a different reference than in its default reading), then the concatenation of p and t should not accept all the complements which t accepts. Let us look at an example which illustrates this claim: Zeno Vendler (1967) prominently argued that some of the terms he calls perfect nominals, terms such as ‘the collapse of the Germans’, are ambiguous; while in their default reading they denote events, they can be combined with some operators, such as ‘is a fact’, which shift their reference to a fact. His thesis is well supported by the observation that there are many complements a perfect nominal normally accepts, which are not accepted by the term ‘fact’. 14 ‘Properties and Kinds of Tropes’, pp. 28f. 15 Surprisingly, Moltmann thinks that, contrary to the predicate ‘is a property’, the predicate ‘is a quality’ is not reference shifting (see ‘Properties and Kinds of Tropes’, pp. 24f.). However, English dictionaries treat ‘quality’ and ‘property’ as interchangeable; the OED, for instance, characterises ‘property’ as follows: ‘An attribute or quality belonging to a thing or a person.’ There may, of course, nevertheless be certain minor differences between the usage of ‘property’ and ‘quality’, but it is doubtful that these could entail any ontological distinction. Anyway, to avoid these issues, I will henceforth focus on uses of ‘property’ (I shall briefly come back to this issue in section 4.b.). The Importance of ‘Being Earnest’ Page 10 / 17 Accordingly, we can observe the failure of certain inferences, which would be valid if no reference shift had been invoked. If, for instance, the collapse of the Germans was gradual and bloody, it does not follow from the truth of ‘The collapse of the Germans is a fact’ that there was a gradual and bloody fact. However, things are different in the present case. If Moltmann’s (M5) was correct, we should expect that the alleged reference-shifting contexts, such as ‘is a property’, do not combine with expressions such as ‘nice’, ‘rare’, ‘peculiar’ etc., which combine with abstract nouns in their default reading; or that such contexts at least trigger significantly different readings of these expressions, which would parallel the differences between (9) and (9*). But contrary to what should be predicted on the grounds of Moltmann’s thesis, these phrases combine well with the alleged referenceshifters, and without change in sense. Indeed, the categorial term ‘property’ is very often combined with such adjectives as ‘nice’, ‘rare’, peculiar’ etc. To wit, take a look at the following quotations: [E]very package belonging to that lady had the inconvenient property of requiring to be put in a boot by itself [...]. (Charles Dickens: The Life and Adventures of Martin Chuzzlewit) There is nothing in which our Species so far surpasses all others, than in the Capacity of acquiring the Faculty of Thinking and Speaking well: That this is a peculiar Property belonging to our Nature is very certain [...]. (Bernard de Mandeville: The Fable of the Bees) Then crush this herb into Lysander’s eye whose liquor hath this virtuous property, to take from thence all error with his might, and make his eyeballs roll with wonted sight. (William Shakespeare: A Midsummer Night’s Dream) He recognized me with evident pleasure; for my rare properties as a patient listener invariably make me a favorite with elderly gentlemen and ladies, of narrative propensities. (Nathaniel Hawthorne: Twice-Told Tales) These quotations show that properties (even if explicitly specified by gerunds or ‘to’plus-infinitives) can be said to be inconvenient, peculiar, and virtuous. Now certainly, in the quoted contexts these latter phrases do not trigger any kind of deviant reading but have exactly the same reading as in ‘earnestness is inconvenient’ etc. Furthermore, we see that, pace Moltmann, the instance-distributing predicate ‘is rare’ is after all combinable with the term ‘property’. So, you can say about properties that they are rare. Now, our use of phrases such as ‘nice property’ licenses all kinds of inferences which Moltmann, on the basis of (M5), should expect to be invalid: whoever asserts 16 Cp. Z. Vendler, ‘Facts and Events’, in his Linguistics in Philosophy (Cornell UP, 1967), pp. 122–146, at p. 142. The Importance of ‘Being Earnest’ Page 11 / 17 ‘Socrates had the nice property of wisdom,’ is obviously committed to the claim that wisdom is nice; from ‘Socrates had wisdom’ and ‘Wisdom is nice’ we are, on the other hand, entitled to infer ‘Socrates had a nice property.’ Similarly, if wisdom is desirable, then from ‘Wisdom is a property’ it follows, on any ordinary standards, that there is a desirable property, and equally it then follows from ‘Wisdom was the virtue that Socrates was most famous for’ that the virtue that Socrates was most famous for is desirable. All this shows a deep misconception in Moltmann’s account: what we say about the referents of abstract nouns can, in the very same sense, be said about properties. The alleged difference between properties and kinds of tropes, consisting in the circumstance that some predicates only apply in different ways to them while others apply only to the latter, breaks down in light of the above quotations. The point I made can be illustrated with the kind of sentence on which Moltmann primarily relies: (13) Shyness is a nice property. is not only perfectly acceptable, but in its natural reading it amounts to whatever the natural reading of (9), ‘Shyness is nice’, amounts to; indeed, (13) is ordinarily recognised as nothing but a stylistic variant of (9). But then, no reference shift has been invoked by ‘is a property’; in the context of (13), ‘shyness’ refers to whatever it refers to in (9): viz. shyness. And since this is correctly said to be a nice property, it is correctly said to be a property. 4. Revisiting Moltmann’s Data a. Aspectival Predicates Hitherto, I have argued that Moltmann’s denial of the Uniformity Thesis is a failure. I presented a direct argument to the conclusion that shyness is identical to the property of being shy; Moltmann’s retreat to (M5), the claim that ‘is a property’ generates a reference-shifting context, has proven unsatisfactory, because it does not square with the possibility of combining the term ‘property’ with exactly those adjectives that Moltmann holds cannot characterise properties (at least not in their ordinary understanding). Moltmann’s crucial theses (M1) and (M5) are mistaken. But yet, something remains to be explained: the linguistic observations with which Moltmann motivated her proposal. After all, there certainly is a difference between (9) Shyness is nice, The Importance of ‘Being Earnest’