Conversion or alternate class membership? Comments on Brøndal’s theory of proper noun*
In the following pages I will be reviewing Br ndal’s (1948) theory of the proper noun, more speci cally, I will discuss some problems concerning what is usually treated in traditional grammars under the heading \reclassi cation of proper nouns (PN) as common nouns (CN).“ Br ndal’s views on word classes, in general, are very intriguing and also extremely thought provoking as well as fairly idiosyncratic and obscure in some respect; at rst sight, however, some of his views seem counterintuitive or simply absurd. I will also examine Curme’s (1935), John Stuart Mill’s (1949), Jespersen’s (1924/1992) and Langacker’s (1991) view on the proper noun, concentrating on how they analyse and comment on cases which are usually treated as PN CN conversions. In various grammars there is a section which is devoted to the analysis of structures in which the name of a well-known person is not used to refer to the person himself but denotes some quality or characteristic associated with that person. Such occurrences of the proper name are analysed in these grammars as reclassi cation, or use, of the proper noun as common noun. In the following sections I wish to examine and comment on this claim suggesting that the traditional subcategorisation of the noun class into proper nouns and common nouns is untenable since the semantic properties of nouns do not parallel their syntactic and/or morphological characteristics, and grammar can only manipulate syntactic information. §0.
!
In a section of his book on the parts of speech 1 Br ndal discusses and comments on the earlier de nitions of the proper noun (pp. 57{63). He §1.
- The idea of this essay has grown out of a class which discussed some traditional
grammatical concepts. 1
I had access to the French translation. See details in the References.
T H E E V E N Y E A R B O O K 5 (2002) 77–94
ISSN 1218–8808
77
78
Lajos Maros an
agrees with the ancient grammarians in that the de niton of PN is impossible on morphological basis since proper nouns combine with the same axes as common nouns. Therefore, in the rst grammars the morphological de nition of \noun” was complemented with a logical de nition since its characterisation as \word with case“ | ptotikon | was not distinctive enough: participles and also pronouns in ect for case in Latin and Greek. The reason why appellatives (that is, common nouns) and nomina propria have been treated as two subgroups of one class is due to the similarities both in morphology and syntactic distribution, hence the diculties of a formal de nition; thus, Br ndal follows the age-old tradition of discussing PNs in logical terms. The key elements of Br ndal’s de nition are: (a) in contrast to a common noun a proper noun stands in an arbitrary relationship with its denotatum,2 that is, there is nothing in the proper noun-word that makes it more applicable to one than to another entity, therefore, (b) proper nouns have no meaning, and are © morphologically simple. Further, he questions Donatus’ position | also embraced by other grammarians | who claims that the PN is the name of one individual whereas the CN is that of many individuals or a class. Br ndal argues that Christian names, such as Pierre or Jean have been the names of innumerable individuals so this de nition, that is, one that invokes the idea of uniqueness, is too restrictive. Still, he maintains that there are some CNs, for instance le meunier, la plage, le bois, which, though they have the potential to be class names, are also used to denote de nite individuals, not a class in certain contexts; in other words, Br ndal also embraces the notion of unique reference in some cases. It is fairly obvious that the two criteria, that is, (i) arbitrary relationship between the word and its denotatum and (ii) uniqueness, involve two di erent aspects of reference. A common noun, or rather, a noun phrase headed by a common noun can have unique reference, too, in a particular universe of discourse, as in Put the fridge in the corner. While the criterion \arbitrariness” (or: non-arbitrariness, in case of common nouns) emphasizes the relationship between a word and its 2
“Arbitrariness” here means that the proper noun word — in contrast to the common nound word — has no abstract description which enables the speaker to apply the word to one rather than to another entity. The other meaning of arbitrariness: the lack of inherent relationship between a linguistic sign and a particular meaning is irrelevant here. See details below.
Conversion or alternate class membership?
79
denotatum, the notion of \uniqueness“ relates the word to a particular universe of discourse. Otherwise, Br ndal nds the application of the notion of unique reference with respect to proper nouns unsatisfactory, therefore, he is critical of the \statistical view” propounded by some grammarians who contrast proper nouns with common nouns proposing that PNs are the names of few entities whereas CNs are applied to many objects, (which can be taken to be a sloppy way of expressing the view that PNs have unique denotation). It is not hard to agree with Br ndal in general: if one adopts a logical approach, what counts is the relationship between the entity and the noun associated with it rather than the number of entities which could potentially be denoted by a particular word. This section examines Curme’s position on the proper noun. Curme’s approach to proper noun is similar to that of Br ndal; unlike the latter, however, Curme is more generous as far as illustrative examples and explanations are concerned. §2.
(1) I never knew a Cummings to stand in the way of progress (2) He is a Cummings through and through Curme (1935 : 1 ) suggests that a Cummings is a proper noun in (1) whereas the same phrase, or only just the noun, is a common noun in (2). It is clear that logical analysis is the basis of Curme’s distinction, a distinction which grammatical considerations do not warrant: identical forms sit in (more or less) identical positions. (3) The Cummingses have left town for their summer home (4) You will nd Cummingses active in the various benevolent activities of our city (5) I never knew a Cummings to stand in the way of progress (6) The Cummingses will give a reception this evening3 3
At this point Anonymous Reviewer asks: “Isn’t belonging to the Cummings family a common characteristics the Cummingses share? I think the denotatum of the word Cummingses does not include people incidentally called ‘Cummings’ but not belonging to the family in question.” I suppose this is a comment on the claim that proper nouns are in an arbitrary relationship to their denotata. The arbitrariness criterion, however, may remain valid even if we complement it with pragmatic factors as to how entities acquire their PNs. See also §7.
80
Lajos Maros an
(7) Christians shouldn’t do such things Curme analyses the occurrences of the name in (3), (4) and (5) as PNs and explains that the plural marker does not make the word into a CN. The expression the Cummingses denotes a particular group in its entirety | hence the de nite article and the plural marker |, the members of which are each single in kind, therefore, are not marked by any common characteristics. The subject expression in (7), however, is di erent: Curme explains that in this case the idea of class enters into this plural, that is, when the division is made on the basis of some common characteristic(s), the article is dropped and the structure in question is not a proper but a common noun. He claims, then, that an item is a common noun if the entities to which we can apply 4 the particular noun have some common characteristics|hence the idea of class |, whereas the individuals to which we refer to with the help of a proper noun cannot be distinguished by any common characteristics. In other words, as pointed out above, proper nouns stand in an arbitrary relationship to their denotata, therefore, their use is unmotivated. To further illustrate this issue Curme explains (1935 : 2) that, though the rich and the poor “represent distinct groups, [. . . ] they are not particular groups, for the members of each group are gathered together on the basis of common characteristics.”
In other words, the rich and the poor are common nouns, while the Cummingses in (3) is a proper noun, because the individuals belonging to
the group do not share any common characteristic(s). The other aspect, which is characteristic of Curme’s (Br ndal’s and other grammarians’) approach to proper nouns, appears implicitly: he never actually discusses the problem from a grammatical point view, that is, how much stretch of speech he considers a proper noun, or more correctly, a noun. Does Curme want to tell us that these noun phrases, the rich and the poor (i.e., det + word), are nouns in this form? Does the presence of the de nite article change these adjectives | rich and poor | into nouns, in particular, common nouns? He never tells us whether the de nite article is part of the \nounhood“ of these expressions. Shall we consider a two-word expression, such as the Cummingses, a proper noun the same way as, for example, the single word John ? Also, \the article is dropped”: does this statement simply 4
Curme claims that a proper noun is the name of an entity, whereas a common noun is applied to an entity.
Conversion or alternate class membership?
81
mean that in contrast to the other forms this NP does not contain a de nite article; or, are we to understand that the word form lexically possesses an article, such as The Hague, which could or should be dropped in certain denable contexts? The diculty of de ning what a proper noun is|as long as we assume there is such a category|lies basically in the fact that neither a purely logical | the more traditional approach | nor a purely grammatical characterization | adopted by structuralist grammars | suces. 5 The logical approach ignores or, at least, is not explicit enough about, the size of the linguistic units it wishes to see as proper nouns; for instance, Br ndal does not make it explicit whether the proper noun is the whole phrase in le meunier, tou Isaak, le Detroit de Bering or just a part of it. On the other hand, more modern approaches adopt the category label PN but apply it to items which need not necessarily be labelled so (for instance, in Hungarian; see below). Grammatical speculations and tests make cases identical which we think are di erent, maybe, only due to tradition and education. To summarize the section: Curme claims that a distinction should be made between common nouns and proper nouns on the basis of whether the structure in question describes some characteristic or not, respectively, which is one of Br ndal’s claims, too. As the examples show, the same item can appear in both qualities. The notion of uniqueness, however, is not invoked in Curme. Br ndal criticizes the opinion according to which the PN is the richest, the most saturated and most specialized in content; Br ndal’s criticism is coherent with the notion that a proper noun has no meaning. This view has occurred every now and then since ancient times and can also be associated with Jespersen in the rst half of the 20th century. Jespersen (1924/1992 : 64{71) explicates his position contrasting his theory to that of John Stuart Mill’s. Mill (1949) contends that common names di er from proper names in that, while the former both denote, that is, pick out entities, and also §3.
5
Anonymous Reviewer suggests that “[t]he two should not be opposed. Proper nouns may behave differently from one language to another (including their phonological, morphological and syntactic behaviour), yet the logical characterisation of ‘proper-noun meaning’ may be legitimate. The fact that certain names refer in this way is probably a universal phenomenon, though it has different grammatical correlates cross-linguistically.” The point of this essay is to show that the traditional PN/CN distinction is grammatically irrelevant.
82
Lajos Maros an
that is, have some description which enables the speaker to apply that word to a particular object, the latter only have denotation.6 Jespersen thinks that|in contrast to Mill|it is proper names which connote the most attributes, therefore, are more specialized than common names, and this fact gives rise to applying them as common names. Therefore, a proper name (may) become(s) associated with the attribute(s) its owner has and, then, the name can be used to characterize, for example, another person who | in the speakers’ opinion | also possesses the same attribute(s). In this way, for instance, well-known people’s names can be used as CNs appearing in various structures. In Jespersen’s view, consequently, PNs denote the fewest entities while they have the richest descriptive content. Br ndal, addressing the same problem, thinks that the richness in content is only virtual: the psychological state of a speaker | le sujet parlant | is in a constant ux determined by the circumstances and, therefore, di erent ideas are associated with the words he uses. These associations, however, do not transcend the individual, consequently, the logical content of a word| which could be extremely poor|and the associations surrounding it should be separated since these associations are not part of the norm. 7 Br ndal’s purpose is clear: he wishes to keep the logical content of words|whatever it may be | separate from the associations cooccurring with them in each individual speaker. The common denominator of Br ndal’s and Jespersen’s position is that both scholars admit that there is more to PNs than Mill says. Jespersen, exploiting Mill’s term, dubs this \extra“: connotation, whereas Br ndal uses the term psychological associations. It should be borne in mind, however, that the two terms have di erent status in the respective scholars’ theory. What Mill and Jespersen identify as connotation can be thought to be the same as Br ndal’s logical or descriptive 8 content, something a common noun has. In other words, Jespersen, but not Mill, claims that PNs are very much like common nouns | at least, in the sense that they also connote. Br ndal’s position is di erent: he thinks that both common and proper nouns possess psychological associations but connote,
6
This is Brøndal’s and Curme’s view, too, expressed in a different way. That proper nouns have only denotation can be taken to mean that a name is in arbitrary relationship to the entity it refers to, it does not have meaning, and it does not invoke the notion of a class.
7
“. . . un mot peut être psychologiquement très riche, tout en étant excessivement pauvre pour la langue elle-même, c’est à dire pour la norme” (Brøndal 1948 : 62)
8
Brøndal does not contrast logical/descriptive content to denotation as Mill does. However, he does talk about the denotation of nouns.
Conversion or alternate class membership?
83
only common nouns have logical contents. Jespersen’s speci cness of PNs does not occur as part of Br ndal’s theory. Though Br ndal is not explicit on this point (either), yet the only sense I can see in his argument is if one assumes that these psychological associations become | in some way | descriptive or logical content and these serve as the basis of reclassi cation or conversion. I feel the basic di erence lies in the fact that Jespersen, using the term connotation, endows PNs|in Br ndal’s terminology|with a normative content (that is, a content related to the norm), too, while Br ndal contends | and this is fairly explicit | the notions and ideas speakers link to proper nouns are of a more individual and, therefore, elusive | that is, non-normative | nature. Hence the term psychological associations. 9 As already mentioned in the previous section, one might surmise that it is these psychological associations that enable PNs to occur as CNs on Br ndal’s view; these psychological associations \turn into” a descriptive or logical content. The term psychological associations may be misleading nowadays; in my interpretation they refer to the knowledge an individual may have about a famous person. According to long-standing tradition, which is also embraced by Curme, Jespersen and Br ndal, some widelyknown proper nouns, or more precisely, the names of some widely-known persons, cities etc. occurring in certain contexts should be analysed as common nouns. For example, Jespersen’s illustrations are §4.
(8) Brussels sprouts (9) Gladstone administration Br ndal does not deny that some PNs can reclassify as common nouns since a PN can also acquire descriptive content and become a CN. Yet, he never tells the reader explicitly whether the acquisition of this descriptive content is anyhow related to the psychological associations occurring with these proper nouns. Br ndal suggests the PN CN conversions involve mostly technical terms related to a person in some way, such as
!
(10) macadam, ohm, ampere, maillechort 9
To make sure that no misunderstanding arises from terminology: Jespersen’s connotation is Brøndal’s logical content on the one hand, and psychological association, on the other. In Jespersen’s view both CNs and PNs have connotations, while in Brøndal’s opinion PNs have no logical content but possess psychological associations.
84
Lajos Maros an
He, however, distinguishes the items belonging to this case from some other PNs; for instance, (11) C’est un Hercule/un (autre) Platon/un Napoleon. Br ndal explains that these cases, contrary to the position taken by some grammarians, do not count as conversions from PN to CN because, though these names may associate characteristics and qualities in the speakers, the actual content of these associations will vary from person to person, depending on the person’s education, family background, disposition etc. Since the descriptive content, or rather: the content of the associations related to the PNs in question is so elusive, we cannot identify them as converted PNs. So, it seems that macadam is di erent from Hercule in that macadam has a more or less reliable descriptive content while that of Hercule is too vague; more precisely, Hercule does not at all have descriptive content only psychological associations. Br ndal is also silent on the issue how he means that the names in (11) should be analysed as proper nouns. In one section he claims that a proper noun only denotes an entity, that is, it picks out| say | a person. If a clause predicates of a person that he is un Hercule `a Hercules’, in what sense should the contention be interpreted that Hercule is a PN? Hercule is certainly not the man’s name; the clause does not claim that the referent of the subject is Hercules himself, the ancient Greek athlete, who is denoted by the word Hercule. What other interpretation is possible of Br ndal’s statement that Hercule is a PN? Br ndal states (pp. 91{95), following Mill (and the obscure Bertelsen quoted by both Br ndal and Jespersen), that proper nouns can denote all sorts of entities without describing them: people, peoples, families, animals, ships, machines etc., real or imaginary, on the earth, in the sky or elsewhere. What is more interesting, however, is his detailed enumeration of what does not qualify as a proper noun and why. He contends that as long as the French surname Lefevre, meaning `the smith’ in an earlier variety of French, has the same pronunciation as fevre, the CN, and expresses the same concept, it is not a PN. Furthermore, Lefevre cannot be analysed as a PN as long as the (members of the) family are associated with the trade of a smith. Similarly, a place name, such as the French Le Ch^ateau or Danish Borg cannot be the members of the exclusive club of PNs as long as the castle that originally gave rise to the name still exists or its memory haunts the inhabitants of the locality. Only when people forget about the building or the family gives up its traditional trade can we consider these words as §5.
Conversion or alternate class membership?
85
proper nouns. In other words, Br ndal implicitly suggests that speakers of the same language community \vary“ in their evaluation of the categorial status of a word depending on their knowledge of the world: the people who are familiar with the local history will use a place name, such as Borre, as a common noun, while for the others ignorant of historical facts the same word will be a proper noun.10 Thus, Br ndal introduces the notion of what one might dub alter11 the category membership of the same phonological nate membership : word varies from speaker to speaker, depending on non-linguistic factors. John Stuart Mill, discussing the same problem, insists that it does not matter how well motivated the use of a proper noun is; once it is associated with an entity through the act of naming, a speci c relationship is created between the word and the entity which is its denotatum. It is obvious that there are two contradicting options from which one has to choose: Br ndal opted for the view that whenever one can trace the least motivation for the name in any possible dimension, that is, if the name has once been associated with a particular entity on the basis of the descriptive content of the word, it should be considered a common noun. In contrast to this, Mill, while admitting the possible motivations for using a name in association with a particular entity, emphasizes that from the moment a word gets to be used as a proper noun this relationship acquires an arbitrary nature (or at least, should be considered so). Br ndal claims that alternate membership involves words associated with technical vocabularies, such as oesophage, phosphate. For the initiated these are CNs with well-de nable descriptions while the layman uses them as PNs since there is no corresponding description or logical content for him. As Br ndal puts it (p. 93)
10
If we take Brøndal’s statements seriously, these words are problematic from another aspect: Lefèvre is analysable into a definite article and a noun, thus it is the same type of structure as la tour Eiffel : a phrase; therefore, it cannot qualify since one of Brøndal’s claims is that a PN should be a morphologically simple word, similarly to the members of all the other word classes; so phrases such as la tour Eiffel, compounds like Angle-terre or Cam-bridge are excluded just as derived forms, Ital-ie, (these are compounds and a derived form, respectively, on his analysis!). However, one should not forget that phrases, such as le Détroit de Bering, and morphologically complex words, such as (Danish) Rhinen, are labelled as members of this subclass in another section. (See pp. 57–63.) The same applies to his other example: Le Château.
11
The term alternate membership is my coinage.
86
Lajos Maros an “. . . pour la majorité de la nation ils [i.e., technical words] ont le même caractère non-descriptif que les noms propres.”
Invoking chemical elements, however, is not really convincing because, compared to personal names, such as John, they are basically di erent. For instance, phosphate is an element which has de nable physical, chemical and sensory characteristics. Each occurrence of phosphate|whether or not somebody knows it|possesses (as far as I can tell) the same description| or, some sort of description | while the word John has no description or logical content at all which would enable the speaker to identify the person of that name on any occasion. The question is rather whether the ignorance of a word of one’s native language or of the history of one’s own locality etc. should lead the grammarian to the conclusion that that particular word is a PN for the speaker who is not well-informed enough. Also, a more complicated situation can arise when an expert communicates with an ordinary person: the same word has di erent status for the two interlocutors. To reformulate Br ndal’s position: he (implicitly) contends that the class membership of a word|whether it be a proper or a common noun|alternates, depending on extralinguistic factors, such as the speakers’ knowledge of the world at large, thus, the analysis of a word in this sense is only possible if the linguist sets out and examines each speaker’s knowledge of the relevant section of the world. In this section I will be examining Langacker’s position on the topic since, as it soon turns out, it can be compared to Br ndal’s in more than one point. Also, it illustrates that the issue under discussion is still an intriguing topic for research. Langacker (1991 : II/59) also addresses the problem of distinguishing common and proper nouns. He proposes that in a construction, such as §6.
(12) the Stan Smith who used to play professional tennis Stan Smith receives an interpretation in which it is a common noun due to
the presence of the de nite article and the restrictive relative clause. Such expressions are used, Langacker continues, when there is more than one person of this name in the scope of discourse. Langacker adds (p. 59) that the proper name “. . . acts as a common noun grammatically because it is so treated semantically, i.e., the grammatical behaviour is symptomatic of its meaning.”
Conversion or alternate class membership?
87
This statement suggests that Langacker|at least tacitly|adopts the traditional view|also embraced by Br ndal, Jespersen and Curme|according to which a word has a primary membership in a word class | for whatever reason | and in some speci ed contexts it occurs showing the trappings of another (sub)class. Furthermore, he follows an approach also well-known according to which grammatical analysis is secondary in the sense that there is an extragrammatical notional criterion which is either re ected by the grammar in question or not. For instance, if we consider Hungarian personal names, which make up a semantic subclass of the noun class, their separation from the rest of the nouns as PNs is only motivated from an extra-grammatical point of view because they feature the same morphological and syntactic properties as any old common noun (in my dialect, at least). In other words, the grammar of Hungarian personal nouns is not symptomatic of their meaning.12 In contrast to Hungarian, some authors| for example, Quirk et al. (1985 : 245 ) | claim that in English personal names are morphologically and syntactically di erent from common nouns. When it is used as a common noun, | Langacker goes on | it has the \: : : value `a person named stan smith’.” This last comment, however, relegates the problem to another dimension. Whether or not proper nouns, expressions like Stan Smith or John always have the value `a person named Stan Smith/John’. On page 60, Langacker re nes his position about items which can occur both as proper names and common nouns. It seems that whether Stan Smith is a proper name depends on how the universe of discourse conforms to the cognitive model according to which a proper name refers to a unique entity, which ensures the de niteness of the expression. If there are several persons of the same name in the universe of discourse, then the conditions are present for using it as a common noun; which means that grammatical behaviour is really only \symptomatic of : : : meaning.“ Langacker, too, seems to say that the class membership of a linguistic item depends on the universe of discourse in which it is used. Langacker, at least in the examples he illustrates his statements with, links class membership to phrase level (and presumably, as is the usual practice, to word level) units. 12
The editor of this volume points out that names which are phonologically identical to CNs have different morphology, such as vas ∼ vasat ‘iron-nom.∼acc.’ but Vass ∼ Vasst ’a name-nom.∼acc.’. Also, Hungarian place names do make up a syntactic subclass of nouns in that they cannot combine with the definite article (when unmodified). (i) Beszéltem a Péterrel ‘I talked to Peter’ (ii) Voltam (*a) Debrecenben ‘I was in Debrecen’
88
Lajos Maros an
He chooses \uniqueness” as a criterion from the available options o ered by the numerous works of grammarians and logicians. 13 If uniqueness is the central element of the de nition,14 the application of the same name to more than one individual deprives it of its potential to be unique, hence, it must be analysed as a CN. Again, as in Br ndal, we can nd alternate membership of a particular lexical item; in Langacker, however, alternate membership is dependent on the actual situation in which the discourse takes place rather than on the speaker’s mind, thus, it is more accessible for research. One might, however, entertain the idea that Langacker’s suggestion is the same as Br ndal’s if the speaker is ignorant about the names of the people present in a situation. Is it reasonable to say that his intentions mis re if he wants to use a name as a proper noun but in fact he cannot since there are other people present of the same name? Does anything go wrong in such a situation? In my opinion, uniqueness cannot be part of the de nition of a proper noun since it can be achieved by various grammatical devices or even gestures; I nd more important the type of denotation characteristic of PNs which di ers from that of common nouns, the type of denotation which involves the arbitrary relationship between the word and its referent as explicated in logical analysis. If we make a speci c denotation-type the main criterion, Langacker’s model outlined above is not feasible: even when there are three individuals of the same name, say, Thomas, in the universe of discourse, the way the word refers to the referents will be the same in all the three cases. Langacker is certainly right when he contends that in such cases the uniqueness of the reference is not ensured by the name alone. So, in Langacker’s view, then, there are cases in which Thomas has unique reference and, therefore, is also a proper noun; and others in which its unique reference is ensured by grammatical devices, such as the de nite article and restrictive modi cation, but, then, it is a common noun. In Langacker’s view, if we assume a very limited universe of discourse, all the nouns we use to refer to entities within that universe will be necessarily proper nouns, for|within that universe of discourse|the particular thing is unique, and the noun the speaker uses to refer to it uniquely identi es it at the same 13
Proper nouns were defined in different ages by different scholars for various purposes and the following criteria emerged of which the definitions of the category proper noun have been composed: (i) PNs uniquely refer; (ii) PNs are definite; (iii) PNs are definite descriptions; (iv) PNs are arbitrarily related to their denotata; (v) PNs have no meaning; (vi) PNs have meaning; (vii) PNs are deictic; (viii) PNs presuppose existence, etc.
14
As a proper noun, Stan Smith “incorporates the supposition that there is one individual with that name in the universe of discourse” (p. 59).
Conversion or alternate class membership?
89
time.15 Naturally, in a suciently limited universe of discourse Stan Smith will also be a proper noun. In a much wider universe of discourse, however, proper nouns may not fare any better than common nouns: the same word has to be used to identify di erent entities, that is, the NP the table will not pick out a uniquely identi ed object; some identi cation in the form of restrictive modi cation will have to be added to Stan Smith or Thomas, too, if we want to uniquely indentify the right person. Thus, a proper noun in Langacker’s view is not really a grammatical category but a feature of a given universe of discourse which Langacker sought to express couched in the traditional distinction between CNs and PNs. In general, we might ask whether the category proper noun is needed in any grammar at all. It seems that the traditional logical distinction between common nouns and proper nouns is not re ected in grammar, as the examples above illustrate: both CNs and PNs can combine with the plural marker, the de nite and inde nite article and the other determiners as well; both subclasses admit of restrictive modi cation, thus, it seems that this distinction is not motivated either syntactically or morphologically, therefore, the subclassi cation of the category noun into proper and common subclasses is unreasonable. From a logical point of view, however, the explication of the di erence between PNs and CNs may shed some light on the relationship between words and the world. The problem of conversion or reclassi cation16 has been haunting these pages from the beginning. For a long time, conversion has been a handy label to describe either a phenomenon or a process which purports to explain the obvious fact that some phonologically identical items can appear in di erent functions; more precisely, the same element can occur in various syntactic positions, with various morphology, and, further, grammarians would like to analyse these di erent occurrences of the same phonological word as exponents of di erent word classes or subclasses of the same category. For instance, in some well-de nable syntactic contexts the English word bottle occurs as a noun, in others grammarians analyse it as a verb. Also, it is claimed that quite a few nouns may appear alternatively as countable or uncountable nouns in the same way that some nouns can occur in some context as common nouns and in others as proper nouns (and vice versa). §7.
15
As was already mentioned above, this view is not alien to Brøndal, either, who contends that noun phrases headed by common nouns, such as le meunier, la plage, le bois should be analysed as PNs in some contexts.
16
I am using conversion/convert and reclassification/reclassify, respectively, as synonymous terms.
90
Lajos Maros an
The interesting feature of Br ndal’s approach (and Langacker’s, too) is that in his view the same item can be both at the same time depending on the perspective of the speaker and the hearer. In the following section the discussion will concentrate on this assumption. In Br ndal’s manner, then, one can state that the sentence (13) Her boyfriend is an Einstein17
| just like the Hercule -sentence above |
and the interpretation of the word Einstein as a noun with a descriptive content, that is, as a common noun must be part of the norm. The other \norm“ meaning is the family name, the proper noun, therefore, the sentence is ambiguous. One reading suggests that the girl’s boyfriend possesses the outstanding intellectual abilities which Einstein was famous for, whereas the other that he is a person whose family name is Einstein. Br ndal would possibly agree to this analysis though he certainly would not consider Einstein as a common noun on either interpretation, which is the only reasonable grammatical analysis of the rst interpretation of the clause. We could express this more generally: the syntactic structure X is a Y has two possible interpretations: (i) it ascribes some characteristic(s) denoted by Y to the subject, X; grammatically, this characteristic can be expressed by a common noun represented as Y in the NP `a Y’; (ii) it identi es X as an instance of the entity of the name Y; Y is a PN preceded by the inde nite article. Whether these interpretations (that is, the descriptive and the referential, respectively) are possible depends on the actual words that take up these syntactic positions. The descriptive reading is usually possible if Y is the name of a historical person; otherwise, the interlocutors must have personal knowledge of the person whose name is Y, as shown in (14). (14) Her boyfriend is a John With respect to the discussion in the previous section, this sentence has only one reasonable interpretation: the person referred to as her boyfriend 17
Anonymous Reviewer remarks that “[i]n addition to the two readings given by the author, a third type of reading exists in such sentences: ‘her boyfriend is from the Einstein family (whichever Einstein family is present in the discourse universe)’.”
Conversion or alternate class membership?
91
is such that he is one of those persons whose name is John. The other interpretation is not feasible since the name cannot be taken to suggest a famous personality whose characteristics could serve as basis for characterizing the subject, or more generally, it cannot suggest any characteristic. In Hungarian there are three possible syntactic structures: (i) X Y ; (ii) X egy Y ; (iii) X a(z) Y. (15) A baratja Janos his/her friend John `His/her friend is John’
(16) (*)A baratja egy Janos his/her friend a John `His/her friend is a John’
(17) A baratja a Janos his/her friend the John `His/her friend is John’ `His/her friend is the John’ (15) means that the man who is referred to by the subject a baratja is Janos, that is, a person known (or unknown) for the interlocutors and identi ed by the personal name. Thus, the personal name is a proper noun (and the NP Janos is referential). (17) indicates that the friend in question is a particular person called Janos, who the interlocutors know; to convey this meaning, the de nite article is stressed. Or (17) is synonymous with (15): the name is a PN on both interpretations. (16) is also ambiguous: it either means | structurally | that the name of the boyfriend, who is in the scope of discussion, satis es a certain description putatively associated with the word Janos : in this case the name is a common noun since common nouns possess a description which enables the speaker to apply the word to a particular entity. Thus, if (16) is asterisked, it is done so by virtue of its uninterpretability as a description, not of its grammar: Janos is uninterpretable as a description, therefore, it cannot occur as a common noun. Or, (16) may refer to a person whose rst name is Janos. Curme’s claim that such cases, that is, the second interpretation of (16), should be analysed as occurrences of proper nouns is reasonable. However, we may take Curme’s idea of descriptive content to the extreme and suggest that there is no such thing as proper noun at all as far as personal and family names are concerned. In his view, the proper noun is the name of an entity; but since many people can have the same name, one might take this feature to be their common characteristic. Paradoxically, John is a proper noun, on the one hand, because it is arbitrarily associated with the person; but once somebody is called John, this is a characteristic
92
Lajos Maros an
he shares with other individuals of the same name, thus, John should be analysed as a common noun. The discussion in this section suggests that the potential for the two di erent interpretations is ensured by the syntactic form but whether the actual utterance is really meaningful does not depend on grammar. It is theoretically possible that for some people (16) is just as meaningful as (13) would be for a di erent group of addressees. We can agree with Langacker: the grammar is symptomatic of meaning of the form; but whether the syntactic form has the descriptive or referential meaning depends on pragmatic factors. It has just been mentioned that (16) may have two interpretations on structural grounds but whether it does so depends on non-linguistic factors. What are these? To interpret (16) as a clause which ascribes a particular property to the referent identi ed by the subject NP requires that|at least | the speaker should have a knowledge of the world such that there has been a person called Janos, who the speaker singles out as someone possessing a particular characteristic feature to which he, the speaker, wants to make reference using the word Janos. The speaker’s intention is doomed to failure if his interlocutor does not associate any characteristics with the personal name. Thus, for him this word will not convey a characteristic but the name purports to identify a particular person. In this case, then, Br ndal is right when he assumes that the word class membership of a word depends on the interlocutors’ knowledge of the world, but only in the sense that the \message” does not get through. If we assume Br ndal’s analysis is correct, that is, the same item can occur once as a PN once as CN depending on the knowledge of the interlocutors, where does \wordclass“ reside? I assume that speakers use the language unconsciously and, therefore, are not aware that they actually use word classes; speakers, however, are able to manipulate meanings and it is analysis that imposes its categories on linguistic entities for interpreting the intended meanings. Therefore, the analysis of technical terms in Br ndal’s manner, such as phosphate, oesophage, family names and place names, such as Lefevre and Le Ch^ateau, respectively, now as proper nouns now as common nouns requires a wide range of knowledge which is not connected to language itself and mostly inaccessible for linguistic research. The use of the name of a chemical element, family name or place name as a proper or common noun in Br ndal’s sense has no effect whatsoever on the interpretation of a particular sentence, since it is non-manipulable by the interlocutors. In other words, a teacher’s use of | say | the word phosphate in a classroom setting will not be di erent from that of the schoolchildren as far as linguistic structures are concerned. Also, this raises another interesting question: how much descriptive content (or
Conversion or alternate class membership?
93
how many associated ideas) is (are) necessary for a PN to be able to convert to a CN. For instance, does a chemistry teacher’s knowledge enable him/her to use phosphate as a common noun, or more precisely, does his/her knowledge pass a threshold which enables the analyst to say that phosphate in the teacher’s utterances quali es as a common noun or, is it only used as a common noun by a research fellow of some laboratory? Or, consider a person not knowing anything about Greek mythology uses the word Hercules meaning `an unusually strong man’. For this person the structure in which the word occurs is not ambiguous, while for a professor of Greek it may be. But this is a pragmatic fact. Br ndal’s problem does not only exist as a conversion, it is easy to nd examples within the same subclass. Consider the Hungarian common noun kupleraj. Children rst learn it to mean `disorder’ but only later do they acquire the other meaning `brothel’. Can we infer to the child’s grammar from this fact? I do not think so. And, further, what is the unit of measurement of descriptive content? Obviously, descriptive content is not measurable but if a noun can be meaningfully (in the senses outlined above) used as predicate, then, one might suspect it has descriptive content, that is, interpretation is suggestive of the logical nature of the word. To nish the article, one may state that though Brondal’s stipulations concerning PNs are interesting, they do not lead linguistic analysis anywhere because his cases of reclassi cation do not connect to any other aspect of analysis: they are isolated cases. Neither does Br ndal discuss the pragmatic aspects of this question (though this critical remark sounds anachronistic). In my opinion, the relevance of conversion of PNs lies in the fact that a di erent meaning (i.e., referential descriptive) is conveyed so contrary to Br ndal the Hercule -case should also be considered reclassi cation; especially so, since there might be people who have the \right” associations with the word | in the same way as experts’ use of technical terms is considered to be occurrences of common nouns. Thus, Br ndal is inconsistent. To wind up speculations about Br ndal’s theory of proper noun, we may remark that knowledge (or ignorance) of technical terms raises the question of a parallel with the knowledge of one’s own language: how would Br ndal evaluate a speaker’s ignorance of, for instance, adjectives? There are two subclasses of noun: common and proper, and using the former one relies on a description while using the latter is only possible after the speaker has learned that this or that entity should be referred to with this word. This may lead to the claim that an unknown noun is a PN in the particular speaker’s language. But what about adjectives, adverbs, verbs and prepositions unknown to a speaker? Are they also proper nouns? Or rather \proper adjectives“, \proper adverbs” and \proper verbs“?
!
94
Lajos Maros an
In this article we discussed some aspects of the traditional proper noun{ common noun distinction. We found that this distinction is not relevant for grammar if the logical di erence between these two subclasses of nouns is not re ected by syntax and/or morphology. It seems, then, the category proper noun is rather a feature of the non-linguistic context than that of grammar; 18 more speci cally, this term can be applied to characterize the speaker or hearer’s knowledge of the world, or the intended referential properties of entities denoted by the speaker, therefore, the cases which are analysed in chapters of various grammars entitled \reclassi cation of proper nouns as common nouns” are mostly interesting for linguistic semantics. Br ndal’s psychological associations, which I interpret as knowledge of the world, are not relevant either in grammar if di erences in individuals’ knowledge about the world are not re ected in grammar; these are pragmatic issues.19 §8.
references
Brøndal, Viggo. 1948. Les parties du discours. Copenhagen: Einar Munksgaard. Curme, George Oliver. 1935. Parts of Speech and Accidence. Boston: Heath. Jespersen, Otto. 1924/1992. The philosophy of grammar. Chicago: The University of Chicago Press. Langacker, Ronald. 1991. Cognitive Linguistics. Stanford, Calif.: Stanford University Press. Mill, John Stuart. 1949. A System of Logic. London & Colchester. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik. 1985. A Comprehensive Grammar of Contemporary English. London: Longman. 18
Anonymous Reviewer: “. . . again, the opposition is not justified.”
19
Anonymous Reviewer insists that there is “a sharp distinction (maybe also in terminology) between the ‘mode of reference’ characteristic for proper nouns, on the one hand, and the grammatical category of proper nouns, on the other. I understand the necessity of this distinction is among the main claims the author wants to make, but this has to be made explicit. In doing so the author should avoid a sterile debate on what the technical term ‘proper noun’ should stand for. It would be better to just introduce two separate terms (e.g., ‘proper noun’ for the grammatical category, in languages that exhibit it at all and, say, ‘proper meaning’ for the reference discussed).” I think the traditional terminology “proper and common noun” based on the logical definition suffices to distinguish between the two types of reference and therefore, I do not think further terminology should be introduced since — as I claim — the category proper noun is not a grammatical category. Also, without “a sterile debate on what a proper noun should stand for” how could one distinguish between the terms suggested by AR “proper noun” and “proper meaning”?
