Eight Issues in the Pragmatic Theory of Properhood

RICHARD COATES University of the West of England, Bristol Fields of research: historical linguistics, etymology, philology of western and northern European languages, theoretical phonology, the history of dialectology, cultural aspects of the history of English, and local history. EIGHT ISSUES IN THE PRAGMATIC THEORY OF PROPERHOOD Aštuonios pragmatinės tikrinių žodžių teorijos problemos ANNOTATION This article develops an approach to the understanding of proper names called The Pragmatic Theory of Properhood (TPTP), which is built on two central ideas: (1) Proper names have no sense (2) The essence of being a proper name is to be found in reference, not in denotation It outlines the elements of TPTP, its premises and consequences, and proceeds to examine a selection of eight of the issues which remain, or which appear to remain. For some years now I have been proposing an approach to the understanding of proper names called The Pragmatic Theory of Properhood (TPTP). Different aspects of it have been presented in a series of papers (Coates 2000, 2005a, 2005b, 2006a, 2006b, 2009, 2011a, 2011b, 2012). The point of the present paper is to outline the elements of the theory, its premises and consequences, and to examine some of the problems which remain, or which appear to remain. It is therefore largely expository rather than argumentative, and the list of references is correspondingly small. The intellectual history of the theory will not be dealt with here, since that was covered thoroughly in Coates (2006a and 2009), but it is good to recall that my biggest debts are to John Stuart Mill (1843), Ruth Barcan KEYWORDS: name theory, properhood, “tag theory” of proper names, reference, semantics, pragmatics. Straipsniai / Articles 119 RICHARD COATES Marcus (1961), and others in the same tradition, for point (1), essentially what is now called the “tag” theory of proper names,1 and to Peter Frederick Strawson (1950) – to some extent – for point (2). For the full history, Coates (2006a) and its very substantial list of references should be consulted. TPTP is built around these two ideas: (3) Proper names have no sense (4) The essence of being a proper name is to be found in reference, not in denotation Some consequences of these two are: (5) Proper names do not form a determinate set (6) Proper names do not fall into logically secure categories (7) Proper names cannot be translated (8) Theorizing the interface between proper and non-proper expressions has neuro-physiological (and psycholinguistic) implications Whilst the following are not strictly consequences of the two key ideas, they support and amplify them: (9) Whilst proper names do not have sense, they are not meaningless (10) The meanings of proper names which can be accessed are ancillary, and to be distinguished from sense (11) The proper names of characters in literary works may be a special case There is one bone of contention which keeps resurfacing in works of theoretical onomastics, viz. that names are either maximally meaningful or minimally meaningful. I suggest that it is not an matter of alternatives, but that: (12) Proper names are both maximally and minimally meaningful and that the detail of the view one ought to take is determined by exactly what one means by meaning. Putting it another way: the supposed issue is a non-issue. Neither of the two key ideas (1) and (2) is new in itself, but I contend that outside of TPTP neither of them has been followed through rigorously, and that their interaction has not been properly understood. Bringing them together into a slogan, we can say that properhood (understood as ‘the state of being a proper name’) is senseless referring. That is important, because many approaches to name theory defend the view that properhood is senseless denoting, a view that I 1 The “tag” theory is generally viewed as being in opposition to Bertrand Russell’s “description” theory and John R. Searle’s “cluster” theory. Since the “tag” theory is widely though not universally accepted, I shall not engage in the detail of that particular debate in this paper. My position extends the “tag” theory to linguistically well-formed expressions which are ambiguous with respect to being names. 120 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood consider incoherent and demonstrably false. The core meaning of a proper name is its referent in the context of its utterance;2 names, once established, are rigid designators (Kripke 1980: 48). Other kinds of meaning, which may be viewed from either the speaker’s or the addressees’s perspective, can be accessed in the act of reference. But these remain ancillary to the basic act. They include, for instance, an expression’s etymology, probabilistic categorization, connotations, and encyclopedic definition. But since none of these need to be accessed in order to perform reference successfully, they do not form part of the definition of properhood. That does not mean that they are unimportant, but their relationship to the key notions needs to be understood precisely, and we shall investigate them below. From now on, the simple term name will be used instead of the fuller proper name. Firstly, the key semantic terms in (1) and (2) need to be defined. The existence of differing terminological traditions in philosophy and linguistics makes this necessary. Reference is the act of picking out an individual referent in a context of utterance; individual means a singleton, i.e. single instance of any type or category, not just an individual person Denotation is the range of potential referents of a word or other expression (its extension), that is, it is an abstraction from reference and must not be confused with it, and it is a notion that applies outside any specific context; that range of potential referents may or may not share any defining feature or features (intension) [The denotation of a name is initially established through acts of reference (among which, for present purposes, I include name-bestowal; on this notion, see further below), but it is a practical consequence of such acts of reference and not a precondition for the success of such acts. Reference is therefore primary and the denotation of a name epiphenomenal upon repeated successful reference achieved through uttering it. Notice that I do not claim that names have no denotation, as some have thought.] Sense is the network of semantic relations in which lexical words and more complex expressions participate, such as synonymy, hyponymy, antonymy, meronymy, conversity, and so on [Sense is distinct from intension. Names cannot not have sense by the above definition, whereas all and only the referents sharing some name may, contingently, share an intension, e.g. just in case all individuals called Barack are male human beings.] My use of these terms is essentially that of Lyons (1977). 2 Such a view is espoused particularly clearly by the philosopher Burge (1973: 438–439). Straipsniai / Articles 121 RICHARD COATES Now let us introduce and develop a simple example, taken at random from the website of the BBC on 22 December 2011. Howard Archer [an economist, RC] said he felt that more action from the Bank of England was likely in the new year. The name consisting of the first two words refers (more accurately, the utterer of the first two words refers by uttering them) to an individual called Howard Archer, who, in the context of utterance, is likely to be uniquely identifiable, either as part of the common ground, or having already been introduced as a discourse referent, or because he is wearing a lapel badge.3 Of course, it is always possible that some addressees may know more than one person with that name, and that is an unavoidable risk which may result in communication failure. Context will usually be enough to decide which person is being referred to, and indeed to decide that a person is being referred to rather than something else which can bear a name, like a dog, a lake, or a sports club, but successful reference can never be guaranteed: only aspired to and attempted. Decontextualized, the name Howard Archer denotes any individual which bears that name; the consequences of this will be assessed in (3’). Those individuals have nothing necessarily in common apart from their name; anything non-random that they might have in common could be due to genetic factors, and then only if the surname had a single origin, which seems unlikely in this case.4 The name can therefore be defined extensionally but not intensionally. Moreover, Howard Archer has no sense. It contains a word which we know to be used as an English surname, but it does not carry the sense of that word, ‘soldier armed with a bow and arrows’. Its etymology is of no importance or help in identifying an individual in a context of utterance; Mr Archer is an economist, not a soldier. Archer, used as a name, is not a synonym of bowman, nor a co-hyponym, with e.g. baker, draper, smith, of trade or occupation. Of course, it is also not a synonym of another surname, Bowman, nor is it a co-hyponym with e.g. Baker, Draper, Smith, of trade or occupation. So if Archer is used to refer successfully to an individual, it does so without resort to sense and only trivially with resort to denotation: it relies on the addressees’ prior knowledge that Archer might denote a range of individuals who bear that name, and that the individual provisionally identified (more closely specified by the given name Howard) is one of them. There is apparently nothing very new there. But if properhood is grounded in reference rather than in denotation, some interesting things follow. I pick up points (3–6) made earlier, in turn. 3 The web-page in question, http://www.bbc.co.uk/news/business-16300104, explains exactly who he is. 4 It has been shown that many surnames, at least in England, can be traced back to a single original bearer (Redmonds 2002), but of course that is not true of all of them. 122 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood (3’) PROPER NAMES DO NOT FORM A DETERMINATE SET There may be expressions whose properhood might be contested. It is unlikely that anyone would contest Neptune, Isabella, Cambodia, Vilnius, and so on: these are prototypical proper names in that by common consent they have no lexical content, i.e. no sense.5 But what about complex expressions that are certainly candidates to be considered as names, especially such geographical expressions as The United States, The West Bank, The North Sea, The Southern Ocean, The Rocky Mountains, and Huang He ‘The Yellow River’, and expressions denoting a single denotatum, like The Pope.6 It is a very interesting matter to decide whether users of these expressions in particular contexts identify their referents by using the lexical meaning (sense) of the words involved. We will leave aside the apparent generic term till (4’); but as regards the specifics, we can be pretty confident that they do not use their apparent senses. We do not need to resort to psycholinguistic speculation about processing to come to a conclusion. We can simply ask linguistically focused questions such as: Is the sentence The Rocky Mountains are not (really) rocky, or The Yellow River is not (really) yellow, a contradiction? I suggest that they are not contradictory. They may be inaccurate, but an untruth is not the same as a contradiction.7 As for the second one, Huang He is so called from the loessy silt it carries in its lower course, but it is said to have been known as (The) Black River in Old Mongolian (Parker 1917: 11, “[T]he Mongols style it the Black River”), from its appearance in its upper course, all of which demonstrates at least that there is no simple relation between names and their apparent lexical content. If these sentences are not contradictory, then the expressions in question must be names. If they are names, we can immediately see that some names and some expressions which are not names are homonymous (The Rocky Mountains versus the rocky mountains, with the name distinguished conventionally in writing by capitalization). If that is true, then there is no fool- 5 They may gain a sense by tropes (rhetorical figures), as in We all hope there will never be another Cambodia (alluding metonymically to the massacres there in the 1970s), but that is irrelevant here. 6 I am referring to the leader of the Roman Catholic Church, though I am aware that other denominations, such as the Coptic Christians, also have a pope. The important point here is that the relevant denominations each have only one person denoted by this expression. If readers think this duality or plurality undermines my argument, they may substitute The Dear Leader, understood to denote the late Kim Jong-il alone; both his father and his son were given different “styles”. 7Cows generally live longer than farmers is factually false, but it is not contradictory. God is three and God is one is a contradiction, but at least some religious people will declare that it is true. Straipsniai / Articles 123 RICHARD COATES proof way of deciding whether an expression, as a linguistic entity, is a name or not; the question can only be addressed in relation to a context. In some context, the expression in question is either used to refer semantically (using the content of the lexis it contains) or onymically (not using the content of the lexis it contains). Properhood is therefore, as illustrated by these examples, a mode of reference, and not a category of expressions, and this understanding of the notion can be applied even to those expressions which are uncontroversially names (e.g. Neptune, Vilnius): they are simply expressions which are never used to refer semantically, except in tropes (cf. footnote 3). An important issue which is in the background here, but which I have not spelt out so far, is that expressions may become names, i.e. they may evolve into names through repeated usage in such a way that users no longer access their lexical content in order to achieve successful reference. This must be true, otherwise normal historical onomastics, which seeks to explain obscure names by relating them to their etymological lexical source, would be impossible and unthinkable. Expressions such as those mentioned in the last paragraph may be names for some users (those who do not access their lexical content), possibly not names for others (those who do access their lexical content), and variably names for yet others (those who sometimes access their lexical content). Such expressions may be on the way to becoming names, if they are not names already. But it is not really an either-or matter. I have argued elsewhere (e.g. Coates 2005b) that once a user has used an expression as a name (i.e. has used it to refer onymically), it is improbable that, under normal circumstances, that user will use it semantically again (allowing perhaps for a brief transitional period). This is because semantic reference is mediated by lexical and grammatical meaning, and its processing costs are therefore higher. Zipf’s Law (or an analogue of it; Zipf 1949) requires us to believe that the processing systems of language-users will apply the least effort required to achieve their referential goal. There is a hypothetical neurophysiological consequence: if onymic reference involves bypassing lexical meaning, there must be a physical neural route which is less costly. This is investigated in (6’) below. Names may therefore evolve from non-onymic expressions in the manner described more fully elsewhere (e.g. Coates 2012). They may also be coined directly out of lexical material and applied to an individual, as when the designation The Former Yugoslav Republic of Macedonia was agreed for certain purposes in 1993 for the new European state. Names may also be reapplied to new individuals. The name Richard was already in existence when it was applied to me. These considerations indicate that names as terms denoting individuals may arise either by evolution or bestowal. The crucial factor in both cases is that the process cancels or renders inapplicable any sense that the expression previously had, either 124 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood gradually or instantly.8 It is important to note that this does not mean the sense cannot ever be recovered during processing, and we return to this matter below in (7’), and (8’). (4’) PROPER NAMES DO NOT FALL INTO LOGICALLY SECURE CATEGORIES Returning to generics of the kinds in the examples introduced in (3’), we can ask a similar kind of question. Names containing a transparent generic will often be the names of what the generic denotes. The Bank of England in my earlier example-sentence is such an instance. But consider The Evening Star is not a star and The West Bank is not a bank. The first of these is true, as normally understood, because the key expression is generally used to refer to Venus, which is a planet, not a star. The second one is true because the expression as a name is generally used to refer to a geopolitical entity in Palestine (though of course it may have other referents too), and there it is metonymic in origin, taking its name from the bank of the river Jordan. These show that expressions which might be considered to be names may appear to contain a generic which does not appropriately categorize the individual. It follows that the sense of that generic is inapplicable or absent when such names are processed. A favourite example of mine concerns an object in my home town, where there used to be a perfectly ordinary railway bridge called Peak’s Tunnel. Peak’s Tunnel is not a tunnel is therefore evidently not contradictory. There is no necessary connection between a name including Tunnel and a thing that is a tunnel. Even if there is a strong expectation that such a relationship will prove to be valid, it does not have logical security. Names containing transparent generics therefore give rise to expectations governed by real-world experiences, but nothing more. A more radical development of this idea is possible: no name carries with it an entailment or a presupposition about the category of the individual bearing it, as many linguists and logicians (e.g. recently Van Langendonck 2007: 71ff. and chapter 3, with references there) have wanted to believe.9 I regard this as self-evident, 8 Let us just follow through the consequence of that: if I name a child using an expression that has lexical content, e.g. Nadežda ‘hope’, that content ceases to be applicable in any act of reference to the child, e.g. “Nadežda is playing with her toys.” This is self-evident. 9 Using Van Langendonck’s terminology: there are no categorical presuppositions. Of course, there are expectations of a statistical kind. Irinas are likely to be female human beings; it follows that Irina, viewed as a name-type or proprial lemma, does not presuppose that the bearer is female. Van Langendonck is right to claim that proprial lemmata yield possible denotata, but the link is not a logically watertight one, as I go on to demonstrate. Straipsniai / Articles 125 RICHARD COATES and cannot understand why anyone continues to believe the opposite. You may wish to categorize Irina as a female given name, or Egypt as a geopolitical name, or Vaiciulenas as a surname. Why do you believe these things? You have experiences derived from encounters (or other ways of acquiring knowledge) that at least one individual bearer can be categorized in this way, i.e. that a particular female, X, has the name Irina, and so on. But it does not follow that an Irina must be a human female (unless the applicability of given names is regulated by law, as it is in some states – not a matter of linguistics or logic). From such a perspective, Shirley was a (rare) male given name in the English-speaking world until Charlotte Brontë used it as the name of the heroine of her novel of that title; it was reinforced as a female name by the fame of the American child film-star Shirley Temple, which sparked off a fashion to apply it to females in the 1930s. Did this name change category from male to female? What should we conclude? At one level it is simply common sense that Irina and Shirley are girl’s names and Vaiciulenas a surname. These classifications are the fruit of repeated observations. They depend on those observations, backed by a type of inference exemplified by If one member of category X is called Irina, then Irina is a name ONLY applied to members of that category. This inference is fallacious: it is an instance of induction generalizing from observed cases, i.e. arguing from the particular to the general, and it is false in the same way as All families have TV sets or All dogs have four legs is false. It is also applied normatively: If one member of category X is called Irina, then Irina is a name which SHOULD only be applied to members of that category. This is a social and cultural norm,10 not a linguistic or logical rule. We cannot escape the conclusion that there is nothing linguistic about Irina that makes it a female given name. We need to distinguish between two claims: Irina is a name applied (only) to individuals which fall into the category of human females and Irina is a member of the linguistic category female personal name The first is true, but if there is an implicit or explicit only it can be falsified by one counterexample (as in the case of the male given name Shirley). Without the only, it can be understood as an expectation or a cultural norm rather than a necessity. It is about tokens of the name Irina, rather than its type. The second means essentially the same as the first with the (implicit or explicit) only. They 10 In either of the senses of the term norm, statistical or coercive. 126 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood make an assertion about Irina as a type. We can conclude that names as tokens can be categorized individually, and classified on the basis of expectation as e.g. hydronyms, theonyms, female given names, and so on. Names as types cannot be so classified. The consequence of that as regards the making of inferences or conveyed meanings is as follows: Irina smokes ten cigarettes a day >> Some female person smokes ten cigarettes a day is a reasonable probabilistic inference based on the life-experiences of the utterer or addressee, which is less probably true just in case there is an Irina who is male, or a squirrel, or a car. It does not have the logical force of an entailment. Names do not fall into linguistic categories in the strict sense, i.e. a bearer of the name-type Irina is not required to be a human female, even if they fall into what we might call expectational categories where a prototypical or typical bearer of the name-token Irina is indeed a human female: a different matter altogether. The categorizations of names which onomasticians use are therefore not linguistic categories (categories involving types) but cultural ones (categories involving tokens) and not linguistic ones at all. In case any reader is not convinced by the argument in the previous paragraph, let them consider the following. It is obvious that names may be duplicated across members of the category to which a bearer belongs: Robert has been applied to male humans and presumably will continue to be so applied. But it may be applied to anything else. I could name not only my son but my cat, my car, my computer, my credit card, or absolutely anything else Robert. Being a name allows application to anything which is culturally capable of bearing a name. That is, of course, a satisfying consequence of our fundamental notion that names have no sense. If you would argue that in all of these hypothetical cases I would apply the “male given name Robert” to the objects mentioned, and that it would still remain a male given name fundamentally, I would of course agree that it still falls into that expectational category. Now consider the possibility of a racehorse named Lucina. Is Lucina a hyponym or horse-name (provisionally considering hyponym to be a valid type-category, as exemplified on the terminological page of the ICOS web-site11)? I did not previously know that Lucina was also the name of a Roman goddess identified with Juno, and associated with childbirth, i.e. a theonym. Historically, therefore, Lucina is a theonym, and one might take the view that the horse-name is an application of the theonym. But suppose I had never come across the goddess. History would count for nothing in my onomasticon, and Lucina’s only place would be in the provisional expectational category of hyponyms. It would not be a theonym at all except 11 www.icosweb.net/index.php/terminology.html, accessed frequently. Straipsniai / Articles 127 RICHARD COATES in some Platonic ideal onomasticon, in some approach to linguistics predicated on the idealized native speaker-hearer.12 This would have nothing, in my view, to tell us about how names work, since TPTP is founded on the notion of reference, that is the contextual picking-out of real individuals in real contexts. It hardly needs to be emphasized that names are subject to “categorial” transfer with great freedom. The onymic categories which onomasticians set up – toponymy, anthroponymy and so on – have no rigid permanent membership. Even those names containing explicit classifying terms have no guaranteed logical relation to the apparently relevant onymic category. Most importantly, the act of commemoration, that is naming some individual after some other individual, need not respect any categorial boundary at all. On 4 August 2011, I collected and analysed the names of all the racehorses running at six British racecourses on that day. I could not etymologize them all, but of those I understood I found that 90 out of 309 (about 29%) carried names which were historically the proper name of some other thing: a person, a place, a mountain, an artistic work, and so on; or had a form that clearly suggested that they were created to conform to a stereotype of one of these categories (e.g. Miss Excel). This suggests the existence of (English) cultural rules such as: Human personal names may be bestowed on horses; or, The form of human personal names is suitable for the names of horses. By means of such permissive rules we might preserve the notion that certain names and even names in particular categories are truly representative of particular categories and can be applied to other categories without becoming indices of the new category; accordingly, that would mean that Miss Excel is not in itself a hyponym or horse-name but, structurally at least, remains a female personal name, like Miss Holly Golightly, the name of a horse which ran at a different meeting. But this tactic has an extremely high theoretical cost. If we concentrate for a moment on hyponymy, or at any rate the actually-recorded names of horses, we will soon discover that absolutely any linguistic material can serve as a horse-name. In the same day’s proceedings we find: Classy Strike (noun phrase at bar-level¹; there were many others of the same structure) Accumulate (verb) Vertueux, Patriotic, Jewelled (adjective) Oh So Spicy, Chilledtothebone (adjective phrase) 12 Of which language or languages? 128 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood Ex Oriente, Beyond Conceit, Avec Moi (prepositional phrase) My Body is a Cage, Iphi intombi [Zulu for ‘Where is the girl?’] (full sentence) Ain’t Talkin’ (verb phrase) Act Your Shoe Size [actually 3 August!], Reset to Fit (both (probably) subjectless sentence serving as an imperative) Hip Hip Hooray, Diddums (exclamation) Only You Maggie, Humor Me Rene (grammatically complex instances including embedded reapplied names apparently serving in a vocative function but referentially obscure) The category of hipponyms is therefore, in principle, the same as the category linguistic expression; indeed, it may even be broader, as there is no requirement on a name consisting of a string of words to be grammatical (Poyle Todream) or even to be in etymologically the same language throughout (Fleurie Lover). The “system” – if anarchy can truly be called a system – of British racehorse names is one of total onymic freedom, and there are no hipponymic types. These considerations reinforce the notion that there can be, and should be, no type-based categorization of names.13 Names may fall instead into temporary expectational or probabilistic categories, though of course such temporary categories may be long-lived. (5’) PROPER NAMES CANNOT BE TRANSLATED This is an apparently wrong-headed conclusion that follows simply from the fundamental notion that names have no sense, i.e. no meaning derivable from any lexical relations in which the expression is embedded. If there is no sense, there is nothing to translate. But readers will note that translation in real-life contexts apparently takes place. Paris is Parijs in Dutch; Finland is Suomi in Finnish; Китай Kitaj is 中国 Zhōngguó in Mandarin; Das Mittelmeer is The Mediterranean (Sea) in English; is Petras not Lithuanian for Pierre, and was Nicolaus Copernicus not the same individual as Mikołaj Kopernik? Is the present Pope not Benedict, Benedikt, Benedetto, Benedictus, etc. as required? It is, of course, possible to define translate and translation in a loose way which subverts the present argument. But these are not translations if this is understood 13 This section is related to the material in Coates (2011b) Strajpsnnai / Articles 129 RICHARD COATES to mean rendering the senses of expressions and their constituents from one language to another. The example Das Mittelmeer >> The Mediterranean (Sea) is evidence for translation under some interpretation of the term having taken place at some point in the past, but it is a translation so heavily disguised that English speakers do not recognize the required Latin etymology without training. The example Китай >> 中国 is not evidence for translation, if by that we mean ‘reproduction of the sense of expressions in one language using expressions of another’, but it is evidence for denotational equivalence. The two names are not lexically comparable – the Russian name has no Russian content whilst the Mandarin one means ‘middle kingdom’ – but they have the same denotatum. Since the notion of denotational equivalence of proper names in usage is therefore required anyway, then, we could push the idea that this accounts for other instances. Paris = Parijs and Benedict = Benedikt = Benedetto = Benedictus are instances of conventionalized equivalence with a long-standing etymological basis. We see the same thing in current English references to the German motor racing driver ‘Michael’ (pronounced in the English way) Schumacher, and this used to be the norm, at least in western Europe. It is not semantic translation. I dwell on the point at perhaps unnecessary length because academic works are still written on “name-translation”, there is publicly available software claiming to translate names, and because the position I outline here has publicly been called self-evidently false. I have suggested above, in (3’), that some expressions may be used to refer, in practice, only onymically (Vilnius, Neptune); in transposing texts from one language to another, these can only be replaced by denotational equivalents (Wilno, Nettuno (perhaps even Poseidon)), not translated. Seeking out knowledge of an opaque name’s origin, and then translating that, is not translating the name. To push the issue to its ludicrous extreme, if we were to substitute He That Striveth With God for the state-name Israel (Hebrew אֶרֶץ יִשְׂרָאֵל (Eretz Yisra’el ) would be to translate the name’s etymology, which is not the same as translating the name – because names are referring expressions, and their etyma are not. I have also suggested in (3’) that the same expression may be used to refer either onymically or semantically (The West Bank, The Rocky Mountains). This latter type includes, in principle, all those which are etymologically transparent. It is only when reference is onymic that translatability is impossible. So long as an expression retains the possibility of being used in semantic reference in principle, then so also is translation possible. We must be careful to specify that where ‘names are translated’, it is their transparent etymology which is accessed for the purpose, and not the onymically-referring expression as such: so with the case of (The Ivory Coast, German Elfenbeinküste, for example. 130 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood Is the difference between semantically and onymically referring expressions in such cases a distinction without a difference? I do not think so. I have suggested (in (3’)) a material basis for the different types of reference which predicts different processing costs for the two types, so, for obvious reasons, I am reluctant to blur the distinction. If I am right, then there ought to be observable consequences, but it is rather like looking for the yeti or gravitational waves. There are hints of their existence, but one would wish for something more solid. In any case, we ought to allow for the possibility, as I did above (3’), of using a transparent name like The Rocky Mountains to refer onymically, whilst at the same time its lexical content is available for other purposes, though not necessarily accessed on every occasion of use. One of the purposes for which it might be available is translation. So-called name-‘translation’ may become conventionalized in a way which is perhaps not seen in ordinary translation. Dicke Bertha, the German name of a First World War artillery piece, is literally ‘fat Bertha’, but she was Big Bertha to English soldiers. Ivan Groznyj might better be rendered something like ‘Ivan the menacing’, but he is Terrible in English, and he is not John as one could, but need not, expect. No-one seems to mind the apparent blunder or shift of emphasis in either case, which suggests that nominating (giving a name to) the referent may be considered more important in “translating names” than respecting the sense of the original expression. Another procedure that I have described in more detail elsewhere (Coates 2006a: 373–376) is in a case where place-names seem to be translated systematically according to their appearance rather than according to their actual lexical value in the source-language: that is, there is wholesale folk-etymology, and it is the fruit of this process which is translated. The original names are not (or are only incidentally) translated according to their scientifically established etymology. An example is the “translation” of Hereford into the language of Irish Gypsies of the early twentieth century as if it were Hairy Foot. These last examples show a desire to have a rendering in the speaker’s language of name-material formulated in a different language. In the case of the German gun, it is not clear that translation has taken place at all; rather, what has happened is the creation of a snappy expression as a name for the alien thing, inspired by, but not necessarily giving an accurate translation of, the original. In the case of Hereford, we have translation, but only mediated indirectly by the current lexicon of the source language. It is not therefore translation of the names themselves, but translation of surrogates for those names, which is in some instances the etymology, and in other instances an expression standing in a more complex relationship with the name in question. Straipsniai / Articles 131 RICHARD COATES (6’) THEORIZING THE INTERFACE BETWEEN PROPER AND NON-PROPER EXPRESSIONS HAS NEUROPHYSIOLOGICAL AND PSYCHOLINGUISTIC IMPLICATIONS Name use is about maximizing the chances of successful referring with the most economical means possible. Names are economical precisely because their reference is unmediated by sense. That must mean that names have lower processing costs than fully articulated semantically referring expressions. It seems reasonable on this basis to hypothesize that any expression which, from the formal point of view, could be either a name or a common expression is preferentially treated as a name. In other articles, I have called this the Onymic Reference Default Principle (ORDP): the principle that the default interpretation of any linguistic string is as a name (i.e. it is free of sense). It is very tempting to speculate that this difference must be embodied neurophysiologically: that proper/onymic reference is predicated on a shorter chain of neurones or more efficiently firing synapses.14 My speculation is that once onymic reference by an expression has been established, i.e. when it has become a name, there is no going back; because a name has lower processing costs, certainly as regards human referents, it confers greater survival value, so there are good behavioural and evolutionary reasons to expect onymic reference to be the default mode of reference. At the individual level, I envisage a short period of overlap at most in which a speaker might use a newly-acquired expression either on the route of semantic reference or on the route of onymic reference for the same referent before the latter becomes the norm.15 The ORDP can be theorized linguistically in the suggestion that whenever an expression is used to refer onymically, it enters the onomasticon of the user,16 which thereby reduces its ability to refer semantically to the same referent on a subsequent occasion. One would expect in principle to find neurolinguistic correlates (perhaps definitions) of this lessened ability: it might be expressed in the properties of neurotransmitters 14 This idea is based on the notion of processing cost as it was first conceptualized in Relevance Theory (Sperber, Wilson 1986). 15 At the societal level, different individuals might use different referential modes for the same expression, or different modes for different proportions of their usage. 16 It should be clear by now that the onomasticon is the set of names available its user, in which each name-form is linked to one or more denotata previously established through acts of reference. These name-forms, understood as proprial lemmata, are categorized expectationally but with no absolutely binding force, such that they are understood as being available for the naming of other individuals in (at least) the same category as the established denotata. 132 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood in synapses. In the absence of appropriate testing, it can be expressed metaphorically as the “greasing” or “lubricating” of the onymic route, whilst, relatively speaking, the semantic route gets “blocked” or “degraded”. Let us turn now to those further points which are not logical consequences of the key notions of TPTP, but which amplify them or flesh them out. (7’) WHILST NAMES DO NOT HAVE SENSE, THEY ARE NOT MEANINGLESS The meaning of any name is established through reference, i.e. the bond established with its referent, without the help of lexical sense, in a particular context of usage. It may generalize, through repeated usage, to the state of having a core meaning, which is its denotation: any denotational, i.e. permanent and rigid, bond which has been established extensionally with its historically established referent(s).17 We might think of its denotation as the stored set of potential referents. Names have other meanings. What these have in common is that they are not sense, and are not definable by the application of the tools of logic in the same way that sense-relations are. Such meanings include: • etymological meanings (which are not the same as senses even when they are transparent, i.e. potentially accessible with different degrees of ease according to the training, or lack of it, of the language user; cf. 5’) • logically non-necessary expectations (as opposed to entailments or presuppositions; cf. 4’) • encyclopedic attributes • connotations Encyclopedic attributes and connotations are available in the following manner. The place Brighton (England), and therefore the name Brighton, has the verifiable encyclopedic attributes of a pebbly beach and veteran car displays on the first Sunday of November, and the public connotations of weekends of marital infidelity and sexual politics, i.e. facts and stereotypes; and also “personal” “meanings”, i.e. those derived from a person’s own life-experiences and judgements, such as (for myself) sunny childhood outings, an unusual system for allocating secondary school places (in 2010), or a large population of bulldogs; and the form of its name might suggest an (etymologically misleading) association with the weather-term 17 There are still, regrettably, works on semantics which do not make the important distinction between reference and denotation, or do not make it clearly. Nothing in this article can be understood without this distinction. Straipsniai / Articles 133 RICHARD COATES bright in a way useful for the tourist industry. These features may not be rigorously theoretically distinguishable from each other; my purpose here is simply to illustrate the kind of meaning-“baggage” that names may bring with them, and which may (not must) be accessed during acts of reference and which may (not must) facilitate the intention of successful reference in the context of utterance, a point which applies equally to lexical material. The internal meaning of names is nothing, if all that we mean by that is they have no sense. The corollary is that such meanings as they do possess are all external to themselves: their denotation (the things they could in principle be used to refer to), and their connotations (understandings of the nature of the individuals, e.g. places, that they denote which are dependent on the perception(s) of the user, including common knowledge), as well as the individuals that they are actually used to refer to on particular occasions. (8’) THE MEANINGS OF NAMES CAN BE ACCESSED Let us make fully clear a point raised during the discussion in (5’). I rejected the notion that names have sense, but I may have seemed to reintroduce it by the back door in allowing that the etymology of a name might be accessed for certain purposes, including translation. The concepts are completely distinct. When an expression is being used as an onymically referring expression, i.e. in its function of picking out a referent in a context of usage without the mediation of lexical content (sense), then by definition sense is not accessed. But other aspects of meaning (those in (7’)), conveyed but not asserted, may (not must – that is the important criterion) be in play in particular ways. In picking out a person by name, expectational categories may come into play (4’). If someone tells me to give something to Jane, I am likely to look for a female human being (but of course I cannot rely 100% on the inference involved: Jane may plausibly be a cat). Let us work to dispel any lingering impression that names may have sense, or that etymology and sense amount to the same thing, because we might be tempted to argue that there is a cline or gradient of sensefulness. No-one disputes that there are synchronically opaque names and that these have no sense, for example the English place-names York, Auckland, Ogbourne, Congleton. Any impression of sense-bearing depends on a very non-linguistic property: the etymological sophistication of the language-user. An individual user of these names may acquire, through education, the philological knowledge that the ancestor of bourne once meant ‘stream’, that that of -ton once meant ‘farm, village’, and that land still exists with a sense 134 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood partly comparable with its original Old English one. That knowledge may be treacherous, because the -land in Auckland originally had nothing whatever to do with land (Watts 2002: 9–10), and one would find it difficult to maintain that land synchronically means ‘land’ in this name. The place-name Chelmsford appears more or less transparently to be ‘ford across the (river) Chelmer’, and according to one possible understanding that is what it is, but that understanding has arisen through a sixteenth-century topographer’s intervention, whether mistaken or deliberate, and that understanding does not reflect the “sense” of the elements out of which the name was originally constructed because the river-name has been carved out of the place-name by backformation. Any argument that etymological transparency is a surrogate for some aspects of sense founders on examples like these, and it is simply false that etymological transparency is the same as sense. What about names with an undeniably transparent etymology? Can it be denied that such place-names as Longbridge, Broadstairs, Newtown, and Redhill, or the surnames Butcher and Jackson, or the female given names April and Violet, have sense? Indeed it can, even here where the transparent forms are etymologically trustworthy. Let us work through one such example: as a functioning place-name, Longbridge entails (i.e. logically requires) nothing, not even the existence of a bridge. Given a set of usual assumptions about names, it must entail the existence of a long bridge at the historical moment of naming, but not thereafter (i.e. the ORDP has applied (cf. 6’)). But the bridge’s continued existence or its disappearance is of no linguistic relevance at all. That does not mean that a name like Longbridge is incapable of conveying any meaning at all. Given normal assumptions about how places get their names,18 it conveys information about the past: that there was a bridge there and that maybe there still is, but that inference does not amount to synchronic linguistically-based knowledge of the place. Going to Longbridge does not allow you to believe with the logical security afforded by the senses of the words apparently involved that you will find a bridge there. Some sort of expectation is admissible, but it can be abandoned without a logical disaster in the face of evidence which undermines it: the lack of a bridge. In the face of this kind of evidence, we can conclude that any apparent sense in names is not really sense but (correct or incorrect) etymological understanding, and they are not the same. They are kept apart by the logical security or insecurity of the inferences which they allow. A or the long bridge is what is perceived by the senses, whilst Longbridge does not necessarily reveal anything comparable. In names, a transparent etymology amounts in linguistic terms to a conveyed mean- 18 There are of course now other, arbitrary, ways for places to get names, exemplified by a good deal of the practice observed in America. Straipsniai / Articles 135 RICHARD COATES ing, not an asserted one: i.e. a meaning with no secure logical status, or one which is merely suggested. On the basis of examples like this, we should conclude that “What does a name mean?” is a misleading question, if it is intended to be understood synchronically. “How does a name mean?” is a better one, and that is the only question of synchronic linguistic importance. It is to be understood as “How does a name identify a referent in a context of use?” The answer is “Through shared knowledge of an arbitrary but stable and rigid association of a linguistic form with an individual”. The corresponding diachronic question is: “What did a name mean?” (i.e. “What sense did it have?”) – note that, crucially, the verb is in the past tense, and it acknowledges the disappearance of sense in the transition to being a name. (9’) NAMES IN LITERARY WORKS MAY BE A SPECIAL CASE There is much of interest that might be said about names in works of art in general, and especially “cratylic” charactonyms in literature (Barton 1990). These are names whose form appears to carry some significance within the work in question. Many charactonyms illustrate what we might call The Etymological Onomastic Turn. Such names may be understood with their etymological and arguably therefore their semantic value remaining available whenever they are used to refer to their referent (or perhaps, more realistically, at least on the first occasion). Take for example the names famous in English literature of Ancient Pistol, Doll Tearsheet, Mrs Malaprop, Mr Bumble, Becky Sharp, Gabriel Oak, Titus Groan, and so on. The point of names of this type is precisely to suspend or subvert the general separation of a name from the sense of its component parts; that is what any semantic literary naming actually consists of: the repotentiation or resemanticization of etymology.19 If charactonyms trade on transparency of some kind, it follows from that that they may be translated. To understand this fully, we need to take into account the context of name-bestowal. Literature is art, and it is legitimate to suppose that this fact cancels the normal assumption about bestowal, namely that the act itself cancels the sense of any words which appear in the expression chosen as the name (see 3’ above). Literary name-bestowal, on the other hand, invites the audience or 19 It seems to me there are broadly three types of literary naming: arbitrary (not really a special type at all), cultural (which trades on conveyed meanings) and semantic (which trades on linguistic meanings). The term cratylic may cover aspects of both the last two sorts. 136 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood readership precisely NOT to suspend the connection between usage and etymology, but to perform the complex balancing-act of maintaining both the form of a name and its significance – in some cases even the pre-bestowal senses of the words that constitute it – active for the duration of the literary event, whether it is watching a play or TV program, reading a novel, or whatever. For that reason, in defiance of the argument above in (5’), charactonymic translation is possible (e.g. Dörchen Lakenreisser, in German, literally and etymologically ‘hypocoristic-Dorothee sheet-ripper’ for the tart Doll Tearsheet in Shakespeare’s Henry IV, part II), and such “translation” may be idiomatic and suggestive rather than purely lexical, just like all other translation. That said, it could still be argued that the act of translating charactonyms is exactly like other onymic “translation” and that it therefore accesses the etymology of the name in question rather than any sense which might be detectable in its linguistic form. One does not after all (or one does not have to) investigate a charactonym for significance in one’s mind every time it is used, as one does for the lexical meanings of ordinary words and other expressions in the text, although the possibility is not foreclosed. Finally, we turn to a question which has been addressed from time to time in the literature of onomastics, namely: (10’) ARE NAMES MAXIMALLY OR MINIMALLY MEANINGFUL? This question has repeatedly resurfaced, e.g. as reported in Van Langendonck (2006). It is essentially about whether names, by virtue of being devices for referring to individuals, carry a great deal of information which facilitates individual reference because names approach the state of uniquely denoting an individual (which is a manifestly false assumption, though traceable far back into classical antiquity), or whether they are semantically empty, i.e. have no sense and no wholly determinate range of denotata, as I have argued throughout this paper. The question is not usually appropriately phrased: it embodies a false dichotomy, and it cannot be answered as it stands. A name identifies an individual bearer which may or may not be unique. If that bearer is unique, use of the name carries with it (but subject to the limitations of the user’s knowledge) all the encyclopedic information available about that individual. If the bearer is not unique, the name has more than one denotatum or potential referent, and uttering it is therefore relatively uninformative without assistance from the context of use. Names lack all sense, but as we have seen they may carry a transparent or otherwise learnable etymology, and thereby suggest (not entail or presuppose) category membership through having a salient bearer in that category or through the sheer force of the numbers of bearers Straipsniai / Articles 137 RICHARD COATES in that category. Names may also, of course, carry a great deal of encyclopedic information and they may connote a great deal. To the extent that they efficiently promote successful reference in context, they are maximally informative. But that information is available through ancillary recall processes rather than being encoded in the name itself, and is therefore not necessarily accessed at the moment of utterance. To the extent that they fail to mean through the possession of systematic lexical relations (senses), they are uninformative. The answer to the question as phrased is therefore: “Both”, but it is not really an important one. * * * This article has provided a guide to the foundations and some the consequences of The Pragmatic Theory of Properhood, and shows how some questions which have been considered important in theoretical onomastics might be addressed.20 REFERENCES Barton Anne 1990: The names of comedy. Toronto: University of Toronto Press. Brontë Charlotte 1849: Shirley. Edition ad libitum. Burge Tyler 1973: Reference and proper names. – Journal of Philosophy 70.14, 425–439. Coates Richard 2000: Singular definite expressions with a unique denotatum and the limits of properhood. – Linguistics 38.6 [370 of the continuous series], 1161–1174. Coates Richard 2005a: A new theory of properhood. – Proceedings of the 21st International Congress of Onomastic Sciences, Uppsala, August 19–24, 2002, vol. 1., ed. by Eva Brylla, Mats Wahlberg. Uppsala: SOFI, 125–137. Coates Richard 2005b: A speculative psycholinguistic model ofonymization. – Quaderni Internazionali di Rivista Italiana di Onomastica 1 (Naming the world: from common nouns to proper names), ed. by Dunja Brozović-Rončević, Enzo Caffarelli, 3–13. [Proceedings of the conference “Naming the World: from Common Nouns to Proper Names”, Zadar, Croatia, 1–4 September 2004.] Coates Richard 2006a: Properhood. – Language 82.2, 356–382. Coates Richard 2006b [2011]: Some consequences and critiques of The Pragmatic Theory of Properhood. – Onoma 41, 27–44. Coates Richard 2009: A strictly Millian approach to the definition of the proper name. – Mind and Language 24.4, 433–444. 20 I am extremely grateful for full and constructive comments on this paper by two anonymous referees. 138 Acta Linguistica Lithuanica LXVI Eight Issues in the Pragmatic Theory of Properhood Coates Richard 2011a: Where are the limits of the name? Some remaining issues with The Pragmatic Theory of Properhood. Paper delivered at ICOS-24. Barcelona, Spain; forthcoming in the Proceedings. Coates Richard 2011b: We are surrounded by ony­mies: relations among names, name-types, and terminological categories. Paper delivered at ICOS-24. Barcelona, Spain; forthcoming in the Proceedings. Coates Richard 2012: To bǣre fulan flōde . óf bǣre fulan flode: on becoming a name in Easton and Winchester, Hampshire. – Analysing older English, ed. by David Denison, Ricardo Bermúdez-Otero, Chris McCully, Emma Moore with Ayumi Miura. Cambridge: Cambridge University Press, 28–34. International Council of Onomastic Sciences (ICOS) web-site, Terminology page, www.icosweb.net/index.php/terminology.html, accessed frequently. Kripke Saul 1980: Naming and necessity. Oxford: Basil Blackwell. Lyons John 1977: Semantics, 2 vols. Cambridge: Cambridge University Press. Marcus Ruth Barcan 1961: Modalities and intensional languages. – Synthese 13.4, 303–322. Mill John Stuart 1843: A system of logic, ratiocinative and inductive. London: John W. Parker. Parker Edward H. 1917: China: her history, diplomacy, and commerce, from the earliest times to the present day. New York: Dutton. Redmonds George 2002: Surnames and genealogy: a new approach. Bury: Federation of Family History Societies. Shakespeare William (c.1598): Henry IV, part II. Edition ad libitum. Sperber Dan, Deirdre Wilson 1986: Relevance: communication and cognition. Oxford: Basil Blackwell. Strawson Peter Frederick 1950: On referring. – Mind, new series 59, number 235, 320–344. Van Langendonck Willy 2006 [2011]: Name theory and set theory: the return of the maximum meaningfulness view of names. – Onoma 41, 45–62. Van Langendonck Willy 2007: Theory and typology of proper names. Berlin and New York: de Gruyter. Watts Victor 2002: A dictionary of County Durham place-names. Nottingham: English Place-Name Society. Zipf George K. 1949: Human behavior and the principle of least effort. New York: Addison-Wesley Press. Straipsniai / Articles 139 RICHARD COATES Aštuonios pragmatinės tikrinių žodžių teorijos problemos SANTRAUKA Šiame straipsnyje dėstoma tikrinių žodžių sampratos teorija, vadinama pragmatine tikrinių žodžių teorija. Ji paremta dviem esminėmis idėjomis: 1) tikriniai žodžiai neturi reikšmės; 2) tikrinių žodžių esmė – įvardijimas, o ne denotacija. Straipsnyje trumpai aprašyti pragmatinės tikrinių žodžių teorijos pagrindai, prielaidos bei pasekmės ir nagrinėjamos aštuonios likusios ar tariamai likusios problemos. Įteikta 2012 m. sausio 6 d. RICHARD COATES Bristol Centre for Linguistics University of the West of England Bristol BS16 1QY, UK [email protected] 140 Acta Linguistica Lithuanica LXVI