De macht van het woord
(1988)–D.M. Bakker–
Auteursrechtelijk beschermdEen selectie uit het taalkundig werk van D.M. Bakker
[pagina 38]
| |||||||||||||||
On A-generic Sentences in DutchGa naar voetnoot*0.This paper deals with sentences like
Most probably, these sentences will be read as general statements on kangaroos and triangles rather than as statements on individuals, which is the most natural interpretation of (3):
Sentences like (1,2) are called generic sentences. More specifically, they must be called a-generic sentences, in order to distinguish them from other generic sentences, in which the indefinite article does not occur:
Although generic sentences have much in common, there are also significant differences between sentences like (1,2), (4), (5), (6), and (7). The main purpose of this paper is to elucidate some aspects characteristic of a-generic sentences as compared with generic sentences of other types. The differences between generic sentences will be seen to mirror those between their non-generic counterparts. As an introduction, I shall make some necessary remarks on the noun in general, the mass-count distinction, and definiteness (1.); in 2. I will discuss some recent literature on generic sentences with special attention to a-generic sentences; in 3. I shall expound some of my views on genericity, especially with respect to a-generic sentences. | |||||||||||||||
1.1.There are some general remarks to be made about nouns and noun | |||||||||||||||
[pagina 39]
| |||||||||||||||
phrases before embarking on the problem of generic sentences. By nouns I will understand here common nouns, not proper names. To begin with, let us consider some aspects of the relationship between individuals and countnouns. There is a countless number of uncles in the world, and all are different. This variety is not expressed in the word uncle. In a certain sense, individuality is mirrored in noun phrases like my uncle, John's uncle, uncle Bob, and so on; but to these and even to the most expanded (non-plural)noun phrases there corresponds at least in principle a variety of what I will call ‘individuals’. Individuals thus escape partly what the use of a noun to refer to them reveals of them. Individuals elude nouns in still another respect. To refer to a certain individual, different nouns or noun phrases are appropriate according to the different situations in which a speaker may find himself. In some cases, I may have to refer to a certain person as my uncle, whereas in other cases I may have to call him the doctor. Every noun or noun phrase, if used appropriately, reveals one of the individual's functions; but even an infinite number will not suffice to characterize the individual itself, because nobody will be able to explain why exactly these and no other nouns and noun phrases can be appropriately used to refer to it. Furthermore, the definition of this set - which would reveal in words the individual itself - will have to contain another reference to the individual to whom this set belongs, which, even if a proper name or a running number would be used, presupposes rather than characterizes the individual. Why indeed attach this serial number to an individual, and not that? Apart from that, the definition would contain one of the members of the set it characterizes. Because no noun or noun phrase reveals the individual itself, when we read or hear My uncle is ill it is just by chance that we know which (if any) individual is referred to. We do get, however, some information of another kind. A noun like uncle is typically related to something that is susceptible to change. It is nouns, noun phrases and their equivalents that are capable of being used as a grammatical subject. A predicate like is ill makes sense only if - on whatever grounds - one could expect my uncle not to be ill. At least, that which is referred to by a noun or noun phrase is susceptible to change in that it was not before brought in relation to the predicate is ill in the same sense, by the same person(s) and for the benefit of the same person(s). That which changes must on the other hand remain identical: changes may only be observed in objects that, apart from the aspect or aspects subject to change, remain the same. If a traffic sign showing red light undergoes a change - let us say it now shows green light - then it must be one and the same, because otherwise there would simply be two traffic lights, each showing light of a different colour, and no change. In the same vein, my uncle must remain one and the same person before and after the predicate is ill has been related to it. Thus, what is referred to by a noun or noun phrase is relatively constant. As this property is common to all individuals, the individual as such is not revealed in it. | |||||||||||||||
[pagina 40]
| |||||||||||||||
As has been said above, there is in principle more than one individual that may be referred to by a noun or noun phrase. In so far as some individuals may be called, for instance, my uncle, they are identical. Again, in being capable of being called my uncle, these individuals do not reveal their individuality: it is a property they have in common. Apart from this, we know that every individual partly escapes what the use of a noun or noun phrase reveals of them. What individuals referred to by e.g. my uncle have in common, could then be summed up as follows:
To these a fourth item may be added, viz. a general property of individuals:
In all these respects they are identical. As my uncle, as a phonological form, remains constant with regard to the individuals to which it may refer, I take it that it is closely connected with the remarkable entity defined by the four characteristics. This entity, to be sure, cannot be the meaning of my uncle, because of 3. listed above. On the other hand, it must be sharply distinguished from the individual, as we have seen. It need not be a ‘metaphysical’ entity either: nor a reale which manifests itself in individuals. I think it is something into which an individual will have to be transformed, if I am to speak about it in general terms. Using general terms is a very complicated act. Consider a certain set of aural and visual impressions which I have to identify to call it, say, my uncle. On the one hand, among these impressions there will be such that are typical of one individual; on the other hand, even these will have to be interpreted in the sense that this individual belongs to what is called my uncle. To achieve this I can compare the individual with other individuals I call my uncle. But how could I succeed? Which of the manifold aspects these individuals display simultaneously and successively am I to compare? Is the fact that uncle John wears spectacles really not a reason not to call him so? If not, why not? Or consider the successive visual impressions we have as stages of the moon. In its different quarters, on clear frosty nights, on damp summer evenings, shining through heavy clouds, it manifests itself in a nearly endless variety. How am I to compare these stages to know the individual called moon. Rather, I will check some ‘stage’ by confronting it with a set of features I feel are characteristic of the moon. For checks like these, although they often are carried out rightly, there exists no falsification procedure I know of that is not in fact circular. The same applies for the relevance of the choice of features. Referring to individuals by means of a noun or noun phrase, even a proper name, always | |||||||||||||||
[pagina 41]
| |||||||||||||||
means revealing one of the individual's functions, that is, a set of properties which, according to a speaker's opinion, is a relevant factor in reality: uncle, moon. Although convention is a very considerable force here, individual man is in a way free to construct functions on his own. I can even understand by uncle or moon one function today and another one tomorrow; but I can also restrict the use of uncle or moon far more than is usual in a natural language community, e.g. in scientific discourse. A function, thus being in principle a product of free phantasy, may even be such that in a certain sense nothing corresponds to it in reality. Thus, the unicorn may be said not to exist in a sense different from, say, that of being a mythical phenomenon. But if I speak about the unicorn in such a way, this will only be a significant statement if the unicorn's existence, that is, the existence of at least one individual carrying the function ‘unicorn’, has been assumed. Normally, if a speaker utters a noun or noun phrase, he thereby states his belief that there exist individuals susceptible of being so called. In any case, the speaker says that some person or persons believe so. As not only the function ‘unicorn’, but also ‘my uncle’ and individuals carrying this function are ultimately believed to exist, we may say that the remarkable entity discussed is believed to be part of reality, though not an individual or a set of individuals. | |||||||||||||||
1.2.This entity I will call ‘object’. Externally, an object is an area: attaching the function ‘my uncle’ to any individual(s) involves positing an area to which the function cannot be attached: if every part of the world were my uncle, it would have little sense to use the noun phrase. Thus, differentiation is involved in using a noun, and is therefore possible - but not more - within the object area related to, say, doctor, gold or moon. Now as a matter of fact there are many doctors, there can be this gold or that gold or still other gold; but it is also possible that the object area is not differentiated internally in a way which is considered significant. For instance, the object area connected with moon could be such an area. Differentiation within the object area follows immediately from the fact that objects are related to individuals and mass counterparts of individuals (mci). Normally, man speaks about individuals and mci's. An individual or mci which corresponds to an object is as such a manifestation of the object. The celestial body I can see in the sky is a manifestation of the object called moon, not only because of the nearly endless variety of its stages, but also in that it, as an individual, eludes any noun or noun phrase and that which it has in common with any other individuals. The same holds true for an individual called doctor. In a certain sense, there is only one manifestation of the object moon; there is more than one manifestation of the object doctor. What I call gold equally manifests itself in numerous mci's. Therefore, if a speaker utters a certain noun or noun phrase, he thereby calls attention to, or evokes, one or more individuals or mci's as manifesting an object. Notice, however, that in so far as individuals or | |||||||||||||||
[pagina 42]
| |||||||||||||||
mci's are manifestations of an object, they are identical; and in so far as several sets of individuals are multiple manifestations of, say, the object doctor, these sets are identical. There are two types of internal differentiation of object areas. The first type can be viewed as the result of a random segmentation. The object area called gold, for instance, may be randomly divided into parts (object manifestations) to which the same noun may be applied, and these parts in turn may be divided into parts for which the same holds true. But if you were to divide randomly the object area to which the noun doctor applies, you might get parts which may only be so called in a totum pro parte way. The second type of differentiation thus has to follow certain ‘given’ boundaries between segments. This is because the object area is related to a function which is the result of the ‘cooperation’ of more than one factor, each fulfilling its own function. The object area called moon, for instance, cannot be divided into parts which can be called so. All parts cooperate to form the object. Something comparable applies for the object area called doctor. Whenever there is a manifestation of objects like sun or doctor, this manifestation is either one or more. In contrast, it seems unreasonable even to suppose the same disjunction to be valid for the object gold. True, if you throw a lot of gold coins through my room, I shall certainly count them; but unless I take into consideration the individuality structure of gold coins, or consider them at least as nuggets of a regular form (as happens to be the case), what is thrown in my room can only be measured, not counted. In Dutch, countability is expressed in two ways: firstly, by the pluralization of the noun; secondly, by the combination of an indefinite article and a non-plural noun. A noun like gold has thus a twofold counterpart, e.g. a doctor and doctors. Therefore, if I want to call attention to, or evoke, a certain amount of manifestation of an object, I have only one possibility (viz. gold) if the manifestations are mci's, whereas I am forced to a choice in case the manifestations are individuals: then I must either use, for example, a doctor or doctors. In other words, whereas the referential use of count nouns involves referring either to one or to more than one, in using mass nouns I can refer to any amount. This fact will deserve our attention again in 3.2. | |||||||||||||||
1.3.As we have seen, the effect of realizing a certain noun or noun phrase NPx comes to calling attention to, or more generally, evoking an ‘amount’ of manifestation of the object corresponding to NPx. This holds true both for indefinite and definite NP's. The latter, however, share one more feature. For the description of this I will take as a starting-point an everyday situation. Let us imagine a lamp repairer called Jack. He has just finished the reparation of a floor lamp. By way of precaution, he has removed the shade before and put it on a shelf along with other shades. He says to his apprentice Peter: ‘The shade, please’. The boy chooses a particular shade from the shelf, the one he knows belongs to the lamp his boss | |||||||||||||||
[pagina 43]
| |||||||||||||||
has just repaired. If Jack asked for a shade, any shade of a given set would have done. But because he has asked for the shade, Peter knows that Jack wanted to point to something which has a specific importance. It is the only shade which fits in a given structural whole, viz. the lamp Jack has just repaired. Notice, however, that the same individual could have been called a shade, namely, if it had occurred in a structural whole in which it has no specific relevance. It is, for instance, no more than a shade if considered in a certain connection with the other shades placed on the shelf. As we have seen, an individual is discrete and as such recognizable. This accounts for the possibility to point to it, to pick it out. But from the example above we can see that individuals, although capable of being picked out, sometimes are not. One of the reasons for this is that their discreteness is not considered important, for instance if the shade Jack asked for is seen as not more than one of the shades placed on the shelf. Now this may be because in the given structural whole there are more individuals present that can do the same job. But if I say, ‘Look, a hairpin on the floor’, it may well be that the thing mentioned is the only hairpin extant within the given structural whole. And apart from the fact that it could be called unique in this respect, it would be very easy to set up an argument following which any individual fulfills a specific function within a given structural whole. But as long as I call the individual mentioned a hairpin, the hearer is entitled to infer that according to me any hairpin could have attracted my attention and forced me to attract the attention of others in the same way. For being referred to by a definite NP, however, it does not suffice that an individual is uniquely important. If it is impossible for the listener to find out which individual is meant, it is also inadequate to use a definite NP. Suppose I am standing in the library before a bookcase with a puzzled look in my eyes. One of my colleagues asks me: ‘What are you looking for?’ I: ‘I am looking for a book’. My colleague: ‘In that case, you need not have a difficult time: there are plenty here’. Evidently, my colleague is joking. He aptly exploits a situation in which of two evils I have to choose the least. Although I am looking for an individual which is uniquely relevant for me, I choose not to use the definite NP, because this would be successful only if my colleague would know which book I had in mind. If I had said, ‘I am looking for the book’, I would have said too much. On the other hand, ‘I am looking for a book’ contains too little: it lacks an unambiguous datum from which the listener can infer that the individual referred to is uniquely important. This is the basis for the joke. Why must the listener be able to find out the referent's identity, if the speaker is to use a definite NP successfully? If something is referred to by a definite NP, it is thereby characterized as the only one having an NP-function within a given structural whole. Therefore, the listener will have to use the knowledge of that structural whole as a clue to find the individual referred to. If he does not know the structural whole, he will not | |||||||||||||||
[pagina 44]
| |||||||||||||||
know which individual has been meant. To state the same in other terms, the definite NP refers to an individual pointed to: if I point to something, I cause it to fulfill a unique role in a structured whole. Therefore, it can be said that the definite article (though not having deictic function itself anymore) explicitly characterizes the use of the NP as deictic use. | |||||||||||||||
2.1.Turning now to the problem of generic sentences, let us reconsider
As there is no kangaroo that is not a marsupial, from examples like (1) it could be concluded (compare Lyons (193-197) and see Oomen) that a-generic sentences are equivalent to assertions like All kangaroos are marsupials. This, however, is not true. Consider the following sentence:
This sentence is undoubtedly a-generic and it is true. Of course, there are rabbits that do not live in a hole, but this fact will not falsify (8), as this sentence says that a rabbit lives in a hole by nature and not as a matter of contingent fact (Lyons 195). Thus, (8) differs from All rabbits live in a hole in that it permits ‘exceptions’ and asserts a necessary relation. There are even sentences that assert the necessity of something that is not contingently true, nor logically true:
To account for these facts, Burton-Roberts distinguishes between syntactic and logical analyticity. If (1) shares its deep structure with
then according to him it expresses a linguistic necessity, i.e. to be a marsupial is predicated of the predicate to be a kangaroo. Such a predication of predicates he calls a metapredication. In the case of (1,10), linguistic necessity is matched by logical necessity. As we have seen, this cannot be true of (8,9): the most one can say about rabbits is that, if they would follow their nature, they all would live in holes and that, if all madrigals were popular, it would not be by logical necessity, because nothing becomes popular by logical necessity (or so it seems). Burton-Roberts tries to tie up analyticity (= logical necessity) and linguistic necessity by assuming that, for instance, (9) is conceivable as analytic to a very low degree, and (1) to a high degree. Pure analyticity is found, according to him, in the following sentence: | |||||||||||||||
[pagina 45]
| |||||||||||||||
because ‘it is no non-linguistic necessity that obliges us to admit the truth of <(11)>; it is a wholly linguistic matter’ (Burton-Roberts 182). Sentence (11), however, is not in fact ‘purely’ analytic. The concept ‘side’ is dependent on a distinction between inner and outer space, which distinction presupposes the concept of closed figures like the circle, the ellipse, or, as it is here, the polygon. The concept of triangle can, therefore, be elucidated without using the concept ‘side’:
If (11) and (12) are both true, then each of them at best expresses what a triangle is like in different ways; and therefore neither can be considered the expression of the concept of a triangle. Consequently, the purity of the analyticity of (11) may be doubted. Perhaps no sentences are purely analytic, not even formally analytic ones; but apart from that, if analyticity is a useful philosophical concept, the only relevant distinction between sentences seems to me to be that between sentences which are truly analytic and those which are not. It seems philosophically unrevealing to learn that a certain set of sentences, evidently constructed for linguistic purposes, should be conceivable-as-analytic-in-any-degree. Linguistically speaking, this could only be interesting if the expression of what Burton-Roberts calls ‘linguistic necessity’ had something to do with analyticity. But only the reverse is true. Consider
If (13) were a scientific statement and if it were true, then a field marshal who never carries a gun would not be a soldier. As Burton-Roberts rightly puts it, this is not what is meant by (13). The straightforward conclusion would then be that (13) is not an analytic sentence at all. Burton-Roberts (189-190), however, maintains its analytic character; ‘the utterer of <(13)>, faced with <(14)> as an objection
could reconsider (and decide that he did want to call the Field Marshal a soldier) but he could equally legitimately reply with: “The Field Marshal is not a soldier in the sense in which I was using the term”’. I agree with every word of the passage quoted. But utterances containing such a ‘sloppy’ usage should not be considered scientific statements. Therefore, the question of their being analytic or not simply does not arise. In the following section we shall take a different view, by looking on analytic sentences as special cases of a-genericity. It will turn out that something like | |||||||||||||||
[pagina 46]
| |||||||||||||||
Burton-Roberts' concept of metapredication is indispensable for understanding generic sentences. | |||||||||||||||
2.2.Carlson defends the view that in sentences like
and only in these, there is reference to the kind ‘soldiers’ by means of the kind's proper name, the ‘Bare Plural’ soldiers. Carlson's kind resembles my concept of object very much; what Carlson calls object is that part of an individual which remains the same in whichever events it partakes. Only non-stage predicates can be predicated of this part, in contrast to what is the case with the individual. As an illustration, consider
On first reading, carries a gun seems to be a stage predicate, i.e. a predicate denotating a particular stage in the life of ‘this soldier’, which differs from other stages, in which this individual does not carry a gun, e.g. when he is asleep. There is, however, another reading of (16), namely, that a property of this soldier is mentioned. From several stages in which this individual is seen to be carrying a gun, it is concluded by generalization that carrying a gun is one of his characteristics. In this reading, (16) contains a characterizing predicate. In Carlson's view, in the following Bare Plural generic sentence
reading the predicate as characterizing is the only possibility. This sentence must be seen as formulating the result of a generalization over stage expressing predications, on the basis of which a characterization of the kind is made. Consider now (13) again. This sentence also contains a characterizing predicate. It must also be seen as expressing the result of a generalization over stages. As Carlson takes only Bare Plurals like soldiers to refer to kinds, and as ‘object’ is the only other non-stage argument mentioned by Carlson, it follows that a soldier must refer to an object, and that the sentence is a characterization of this object (cp. Carlson 176). Following Carlson's line of thinking, this seems to be an adequate description of the difference between (17) and (13). It could explain the difference between (18) and (19):
Precisely this phenomenon shows that we shall have to differentiate between stages of objects and stages of kinds. But if we have to, there | |||||||||||||||
[pagina 47]
| |||||||||||||||
remains the question how we are to account for the obvious difference between the object called a soldier (in (13)) and the object called this soldier (in (16), second reading). However generalizing (16) may be, it is always an assertion about one particular soldier, whereas (13) is an assertion about no soldier in particular. Carlson does not give an explanation for this difference. Therefore, his theory is not capable of describing a-generic sentences in so far as they differ from assertions about a particular object like ‘this soldier’. On the other hand, in his theory he has rightly stressed the similarity between expressions like this soldier and a soldier ((16) and (13), respectively). The similarity consists in the fact that in both cases one object and not more than one is involved in the predication. Now whenever such is the case in Dutch, there are in principle two readings, one in which a particular object is meant, and one in which an entity is meant that is, in a way, representative of a kind. Consider the following Dutch sentence:
The individual-reading would force us to the conclusion that the speaker had uttered a useless truism. A more natural interpretation would be captured in ‘a suit like this’ as a paraphrase of dit pak. As (13, 19, 20) show, there are generic sentences containing subjects which will have to be described, in Carlson's own terms, as ‘objects’. In his view, as we have seen, Bare Plurals, although containing the notion ‘more than one’, can refer to kinds, which are one. Carlson, admittedly, adduces sufficient material for establishing similarities between Bare Plurals on the one hand and proper names (Carlson 59 f.) and collective nouns (Carlson 61 f.) on the other, thus establishing a firm basis for the relationship between Bare Plurals and kind. However, according to Carlson there is still another NP, viz. the definite generic non-plural NP, which refers to kinds as well. Carlson states (279) that both NP's denote the property set of a kind, that they both are to be regarded as a proper name of a kind (280), although they each do so in their own way and for that reason cannot be identified one with the other, as Carlson himself admits (274 ff.). Thus, following Carlson, we shall have to distinguish at least three types of generic NP's, namely two referring-to-kind-types and the object-type. It must be said, however, that Carlson does not offer a systematic differentiation. | |||||||||||||||
3.1.The main thesis of this paper is that generic sentences are statements on objects and their manifestations, rather than on individuals and mci's. Every time we call an individual a soldier, we at least say that we recognize something that remains identical in spite of the differences between the individuals so called: a manifestation of something relatively constant, | |||||||||||||||
[pagina 48]
| |||||||||||||||
which in one respect corresponds to soldier and in another respect does not, which is countable and not more than one, and whose discreteness is not considered relevant. Every sentence containing a soldier calls attention to a manifestation of this type. In non-generic sentences, however, the purpose of this is to characterize an individual or mci in order to record its relevance with respect to the state or event mentioned: to achieve this, I have to state what the individual has in common with other individuals, in other words, I have to speak about it in general terms. In contrast, generic sentences are not about individuals, but about what some individuals have in common: an entity abstracted from those individuals, or - which comes to the same thing - an entity constructed as a manifestation of the type described in the preceding lines. By uttering a soldier in this sense, the speaker may be viewed as supposing the entity mentioned to be part of reality, even if no individuals corresponding to it are present at the time. In fact, it is an entity posited by language use, although recognized in every individual deemed worthy of being called a soldier. An a-generic sentence is about such an entity. In so far as the realizing of an NP containing an indefinite article can be seen as a predicate predicating something of such an entity, a generic sentence can be characterized as a metapredication, at least if the subject can be interpreted as such an NP. This concept of metapredication, however, differs from that formulated by Burton-Roberts. As Carlson (175-6) has pointed out, there are sentences like (1) that have no counterpart like (8) and the other way round. Here are some of his examples:
In Dutch, sentences like (8, 16b) sound awkward:
These are two arguments against Burton-Roberts' description of the structure supposedly underlying a-generic sentences: one against accepting it for English and one against accepting it for Dutch. The following sentence which is nearest to (23b) as regards the meaning conveyed, may be considered grammatical, although immediately acceptable to everyone: | |||||||||||||||
[pagina 49]
| |||||||||||||||
This sentence, however, shows a semantic feature not present in (23a). It means something like ‘If it comes to the point, a kangaroo is a marsupial’, or ‘A kangaroo's life is essentially being a marsupial’. It could be that sentences like (23c) are a special type of identity statements like the French sentence ‘Gouverner, c'est prévoir’ and the like. Anyway, (23b and c) typically subordinate a state under a more general type of state, which makes sense only if the listener must be invited to see the dominating marsupial behind the kangaroo. As we have seen, to the object a function corresponds. As this is in principle a product of free phantasy, it is possible that I change the construct overnight. I think this explains the problems around sentences like (8, 13). As a speaker of a natural language, I am free to use a certain function without announcing this by definition or otherwise. If I choose to view the rabbit in the light of a far-off Hole Era rather than in that of contemporary domestication, I can today assert a proposition verbally identical to one I will deny tomorrow. Some would (perhaps rightly) think that this behaviour would be inconsistent; in fact, I think it is only a relatively dramatic variant of an everyday phenomenon. What is decisive here is that as non-scientific speaker I am entitled to include or not to include in that which is common to all rabbits the property of living in holes; this may depend on whether I consider living in a hutch as atypical behaviour, or not. As a-generic sentences are not about individuals, they only have ‘characterizing’ predicates. In Dutch I can say Het dak lekt (‘the roof leaks/is leaking’), when the rain is penetrating through the roof, but also on a bright summer day. In the latter case, the predicate is characterizing (cp. also 2.2. above, sentence (16)). It goes without saying that only individuals and mci's are subject to events; although what is common to individuals may have something like a history, this is not a history of events. Therefore, it is only possible to characterize what individuals have in common in spite of their different stages and their different histories, in other words, what are their relatively permanent properties. More precisely, generic sentences will indicate what are, for an entity, the consequences of being susceptible of being called NPx. If I am right, this is a causal relation. An implication, as Lyons remarks can only be a partial expression for it. And in so far as an implication is a statement about individuals, it is not an adequate expression for the content of a-generic sentences, which are, as we have seen, not about all individuals denoted by, for instance, a kangaroo, but about what these have in common. | |||||||||||||||
3.2.For generic sentences like (18), the same holds true as what is valid for a-generic sentences; they differ in that they characteristically contain | |||||||||||||||
[pagina 50]
| |||||||||||||||
NP's calling attention to what is common, not to single individuals but to sets of individuals. Or in other words, they call attention to more than one manifestation of the object. As we have seen, there is choice in principle between singular and plural in contradistinction to mass manifestations like those called gold, and therefore sentences like (5) may convey roughly the same message as sentences like (4), in Dutch as well as in English. Sometimes, however, what sets have in common will not occur in what single individuals have in common, as is shown by (18, 19), or by (24) a and b:
Because the plural can be used to denote any number of individuals or manifestations except one, it can also denote the total number. Thus, a generic plural-sentence may espress what is common to any set, including the comprising set. In fact, abstraction is made of the differences, including the quantitative differences, between the sets. If something of the kind holds true for English Bare Plural generic sentences, this would explain Carlson's intuitions about their referring to kinds. In this respect, these sentences are much like indefinite non-plurals like (7), and its Dutch counterpart:
This sentence we will contrast with the following generic sentence:
In (26), mention is made of what is common to any occurrence of gold, seen as an entity whose discrete relevance as a structured whole is caused by its being gold. Now, as we remember, the object ‘gold’ is supposed to be part of reality. If it is indicated by a definite NP, then reality is often the structured whole in which the object exercizes a specific function as gold. If this is the case, then only world scale predicates are possible: very general characterizations, which aim at differentiating the object considered from other objects with a kindred function. This could be the reason why the following sentence seems not to be acceptable (cp. Carlson 277):
When we compare angels with beings of other types like animals or humans, we find that playing the harp is not a characteristic of angels, because for instance humans also play the harp. If one thinks of the object called angel | |||||||||||||||
[pagina 51]
| |||||||||||||||
as a part of reality, other things come to mind: not even whether it has wings or not, but its function as a part of the ‘establishment’ of the universe. On the other hand, if you encounter an individual angel, you may find that as a rule it has wings and plays the harp. Therefore, the following sentences will be quite acceptable, I think:
Notice, however, that the choice between definite and indefinite generic NP's may also be determined by discourse. Thus, speaking about pigs, you can begin your speech with the following words:
But after a while, you may continue:
I refrain here from discussing examples of definite plural NP's in sentences like (6). Neither will I go into the so-called attributive use of definite NP's (cp. van Alkemade in this volume). On the latter topic I would permit myself just one remark. If the sentence
is used attributively, then it may be said that the sentence is not about an individual but about what is characteristic of whoever could be suspected of being Smith's murderer. A definite NP is used because the discreteness of the entity evoked by Smith's murderer, within a given structural whole, comes to the fore by being susceptible of being so called, whence the NP can be used deictically. Although much research will have to be done, I hope that these few preliminary remarks will show that there is a possibility for treating the different generic sentences, including the attributive definite NP's, as forming a ‘structured whole’ mirroring the ‘structured whole’ of the ‘normal’, ‘non-parasitic’ sentences containing corresponding NP's. | |||||||||||||||
[pagina 52]
| |||||||||||||||
References
|
|