It has been a long debated question how many articles there are in English. Obviously there are only two material articles, the definite article the and the indefinite article a (an). The distinction thus is between, for instance, the language and a language. However, the noun language, and indeed many other nouns, are also used without any article, as in the sentence Language is a means of communication. It is obvious that the absence of the article in this sentence is in itself a means of showing that "language in general", and not any specific language (such as English, or French, etc.), is meant. Hence we may say that there are three variants: (1) the language, (2) a language, (3) language. Now the question arises, how this third variant is to be treated. The older grammatical tradition described it as "omission of the article", which is obviously inadequate, since there is not the slightest reason to believe that the article in such cases was ever "omitted". Another view is that we should describe this as "absence of the article", and sometimes this notion is made more precise and the phenomenon is called "meaningful absence of article". 1 A third view, which has been gaining ground lately, is that the very absence of the article is a special kind of article, which is then termed "zero article". According to this view, then, there would be three articles in English: definite, indefinite, and zero. This idea of a zero article takes its origin in the notion of "zero morpheme", which has been applied to certain forms in inflected languages, — namely to forms having no ending and differing by this very absence from other forms of the same word, which have each their individual ending. A case in point in Russian is the genitive plural of some nouns (chiefly of the feminine gender), e. g. рук, which is characterised as a special form by the absence of any ending, as distinct from nominative singular рука, genitive singu lar руки, dative plural рукам, etc. The notion of "zero morpheme" may also be applied in English, for instance, to the singular form of nouns (room) as distinct from the plural form with its -s-inflection. If, therefore, we were to interpret the article as a morpheme, the idea of a zero article would make no difficulty. If, on the other hand, we take the article to be a word, the idea of a "zero word" would entail some difficulty. It has been pointed out that the notion of a "zero copula" has been applied to such Russian sentences as он здоров, where there is no verb. In this sentence the present tense is implied as distinct from он был здоров and он будет здоров, where the past or future tense is expressed by a form of the verb быть. However, in this case it is not a "zero copula", but a "zero form" of the copula быть. We might thus formulate the following tense system of this copula: present tense — "zero", past tense был, future tense будет. So even in this particular case the notion of a "zero word" seems very doubtful. Still more doubtful is the notion "zero" with reference to the English article, if the article is a word. We will therefore proceed on the assumption that the notion "zero article" is only possible if the article is not a word. The two main views of the article are, then, these: (1) The article is a word (possibly a separate part of speech) and the collocation "article + noun" is a phrase (if of a peculiar kind). (2) The article is a form element in the system of the noun; it is thus a kind of morpheme, or if a word, an auxiliary word of the same kind as the auxiliary verbs. In that case the phrase "article + noun" is a morphological formation similar to the formation "auxiliary verb + .+ infinitive or participle", which is an analytical form of the verb. 1 Now, the very fact that two such widely divergent views of the article are possible shows that there are some quite peculiar difficulties here. Besides those already mentioned, there is the problem of the meaning of each article: we must find out whether it has one or several meanings, each of them appearing in a different context. We can illustrate this problem by comparing, for example, the two sentences: (1) The dog has come home and (2) The dog is a domestic animal. Of course it is at once obvious that the dog in the former sentence means one individual dog, whereas the dog in the latter sentence means the dog in general, as a zoological species. The question, then, is whether the article itself has two distinct meanings (if so, the second of these is termed "the generic article") or whether the meaning of the article is the same in both sentences, and the difference in meaning between them depends on some other factor. If we endorse the first view, we shall say that the definite article has at least two distinct meanings, viz. (1) it means that an object is singled out from all objects of the same class, (2) it means that the whole class of objects, as distinct from other classes, is referred to. If we endorse the second view, we shall say that the definite article has one meaning only, viz. that of something singled out from other entities. Now, whether the essence thus singled out is a separate object or a whole class depends not on the article at all but on the other elements in the sentence, usually on the predicate. Reverting to the two sentences, (i) The dog has come home and (2) The dog is a domestic animal, we shall see that each of the predicates has several peculiarities which influence the meaning of the sentence one way or another. Let us analyse each of these. First, the grammatical peculiarities. In (1) the predicate is a verb in the present perfect tense, in (2) it is a group "link verb + predicative", and the link verb is in the present tense. That of course would not in itself be sufficient to show the different meanings of the sentences, but it does give a certain indication this way: the verb in the present perfect tense is more likely than not to express a concrete action (i. e. one that has taken place once), while the group "link verb in the present tense + predicative" is very likely to express some general characteristic. Now, these grammatical points are supplemented by some lexical points, which make the difference quite clear. In (1) the verb come and the adverb home denote a concrete physical action and the place to which it is directed, while in (2) the predicative a domestic animal 1 denotes a zoological idea and thus proves that by the dog is meant not an individual dog but the whole species. According to this view, then, the meaning of the definite article itself is the same in both sentences, and the difference proceeds from the peculiarities of the predicates and the words expressing them. Which of the two views is the more convincing one? Both views seem to be defensible, and the decision will have to be made on the ground of some guiding principle. Such a principle may be that of the invariable, i. e. of a stable element in the meaning of a word preserved throughout all the changes and combinations in which the word may be found. This principle of the invariable has been recently very forcefully defended by A. Isachenko in his paper on grammatical meaning.2 The principle may be briefly formulated in this way: "State an invariable wherever possible," or negatively in this way: "Do not state differences wherever this is not strictly necessary." In short, the principle amounts to this. Whenever a word, or a word-form, appears to have different meanings in different contexts, look for that element of its meaning which is always there and does not depend on any context: that is the invariable. If we adhere to this view (as it appears we should) we will say that there is no difference in the meaning of the definite article between the sentences The dog has come home and The dog is a domestic animal; the difference proceeds from other sources, as we have explained above. It is obvious, however, that not everybody will accept the principle of the invariable, and for those who will not do so, the question of the meaning of the definite article will appear in a different light. The same may be said about the indefinite article. If we compare the two sentences, (1) There is a hill behind our house, and (2) A hill is the opposite of a valley, 1 the question will arise, whether the indefinite article with the noun hill has different meanings in the two sentences. If we think it has, we shall say that in (1) it serves to denote an individual object, without reference to its individual peculiarities, and in (2) any object of a given class. If, on the other hand, we endorse the principle of the invariable, the article will be said to have the same meaning of indefiniteness in both sentences, and the difference in meaning will have to be sought elsewhere. We shall first of all note the different types of predicate in the two sentences. In (1) we have the predicate there is,2 in (2) the group "link verb + predicative", and the predicative is a noun. There is, besides, an adverbial modifier in (1) and an object in (2). From the lexical point of view, it is important to note that in (1) we find three words with a meaning pointing to a concrete situation, viz. behind, denoting a relation in space, house, and especially our. In (2), on the other hand, there is the group the opposite of a valley, which expresses some general notion, not restricted to any concrete position in space or time. The indefinite article before valley is of course quite parallel to that before hill, and they are bound to be used in quite the same way. All these peculiarities in (2) point to the sentence having a general meaning, i. e. expressing a definition. Such, then, are the factors on which the general meaning of each sentence and the use of the indefinite article depend. Taking this line, then, we should say that the invariable in the indefinite article is its meaning of taking an object without its individual peculiarities. Whether the noun used with this article is used to denote "a certain hill" or "any hill", is outside the meaning of the article itself, and depends on a series of different factors, which we have tried to point out. It must be emphasised, of course, that if the principle of the invariable is not accepted the result of the analysis will be different.