Studia Linguistica. 1952.

CONCERNING THE SO-CALLED PARTS OF SPEECH


One of the questions in grammatical research most eagerly treated has for a long time been that of the so-called parts of speech. Whoever wishes to gain further knowledge of the long history of this research may with advantage use Ludwig Jeep's treatise "Zur Geschichte der Lehre von den Redetheilen bei den lateinischen Grammatikern'' (Leipzig 1893), Viggo Brøndal's paper on "Les parties du discours" (Copenhagen 1928) and A. W. Groot's essay "Structural Linguistics and Word Classes" in Lingua (vol. I -427-500), all of which contain considerable information concerning this matter.

As is well known, the traditional classification of words into parts of speech has very tenaciously maintained its position in grammatical literature, and without having been in any serious sense shaken by criticism. The form which the classification of 'word classes' received in the Latin grammars (interjections, nouns, verbs, adjectives, numerals, articles, pronouns, prepositions, conjunctions) has in the course of time undergone only minor changes. On the whole, the definitions also remain the same as before. However, criticism of this classification has for a long time been almost murderous. Already Herman Paul says in his classical book "Prinzipien der Sprachgeschichte'' (acc. to 4th ed. 1909, p. 352): "Die übliche Scheidung der Redeteile in den indogermanischen Sprachen — beruht nicht auf konsequent durchgeführten logischen Prinzipien, sie ist vielmehr zu Stande gekommen unter Berücksichtigung sehr verschiedener Verhältnisse. Sie trägt daher den Charakter der Willkürlichkeit an sich. Ihre Mängel lassen sich leicht zeigen." In connection with this Paul also expresses the opinion, which is later found again and again in philological literature, that a better classification is impossible: "Es würde aber nicht möglich sein etwas wesentlich Besseres an die Stelle zu setzen, so lange man darauf ausgeht jedes Wort in eine bestimmte Klasse unterzubringen. Der Versuch, ein streng logisch gegliedertes System aufzustellen, ist überhaupt undurchführbar.'' We can find a somewhat similar conclusion in Erik Ahlman's essay on the definitions of nouns and adjectives in the periodical Virittäjä (1936, ''Substantiivin ja adjektiivin määritelmistä''). Ahlman ends his account by saying: "The fact the committee, (i. e., the Finnish Committee on Grammar, 1915) despite all its study and penetration——— has been unable to achieve a result which is scientifically satisfactory, indicates that it is in fact impossible to define the traditional parts of speech in a way which possesses validity in all cases." This pessimism, however, is, as we know, not shared by all linguists, and repeatedly, and in various places, strenuous attempts have been made to find a more satisfactory solution to these problems.

This pessimism is in itself certainly somewhat peculiar. First, one may reasonably ask, how can a classification be sustained as a general proposition, and how can it be of practical value, if one is unable clearly to define the groups which are being separated? lf, on the other hand, the classification is considered to be of the type, e. g. cars classified into a) yellow cars, b) lorries, c) Fords, it must be within the bounds of possibility to correct it. I would prefer, therefore, to incline to the belief that this pessimistic point of view taken by some linguists is indefensible, and that those who have proposed new schemes for classification into parts of speech are pursuing a road which may certainly lead to valuable results.

But with every classification into parts of speech it is a natural and inevitable assumption—close attention must be paid to the fact that the word "word" is extremely ambiguous. The meaning of the term word is not the same in every case. Any grouping of so-called "words" which does not make this fully clear is of course hopeless. Now if one wishes to produce a classification into parts of speech, the question then arises as to what sense the term word bears; that is, the question of what exactly it is that one is attempting to classify. Personally, I am inclined to think that the difficulties against which research into the question of classification into parts of speech has had to struggle were caused, at least to some extent, by the fact that it was not always clear what was being attempted to be classified.

Without going more thoroughly into the subject I must recall the fact that in philological literature, amongst other things certain visible or audible signs are called words. Such "words" are real things: printed, written, momentarily audible. In a strict, scientific terminology such words may be called just words. I indicate such a word by writing, for instance se 1t, etc. This 1t means "of the first type'. These real things stand in fact on the first level in the logical hierarchy. They represent the lowest, so-called first type. But this is not the only meaning of the term "word", for it can also mean classes of words of the first type; e. g. one calls a word both in Finnish and in Swedish the class of all the graphemes, which contain the letters s and e in this sequence; likewise a word, according to the modern idiom, is the class of all the graphemes which contain the letters s, e and n in this sequence. It is possible to find, for instance, the ”word" with s and e several times on the very same page of a book. Such "words" represent in the hierarchy of conceptions the second type, standing on a higher level. I indicate classes of real words by writing e. g. se 2t, sen 2t etc. These "words of the second type" one may call word-forms. Word-forms in this sense are conceptions, invisible, inaudible and, of course, without any meaning. They do not belong to reality like words of the first type. These cannot be spoken or written down. Thirdly there are classes of certain word-forms, which are often also called "words", but which differ from those mentioned above. They represent the third type in the hierarchy of conceptions. The Finnish grammar-committee of 1915 does not call these "words" "words", but rather vocables. I indicate a vocable by writing e. g. se 3t. The indicated class contains as elements in the Finnish language, inter alia, a nominative se 2t and a genitive sen 2t, in Swedish, inter alia, an infinitive se 2t and the archaic present tense sen 2t. The term "word' means vocable if one states, for example, that the word (the vocable) se 3t has several word-forms, was a nominal inflection, a verbal conjugation.

Does the traditional classification into parts of speech divide into groups the vocables, words of the third type, or the word-forms, words of the second type, or the words, words of the first type?

As far as I can see, classification of the vocables, words of the third type, has been attempted. Thus attempts have been made to obtain a classification which would contain categories for every vocable, with all its word-forms and occurrences of word-forms. The result has been a questionable one, for one thing because it rarely happens that all the word-forms and all the occurrences of word-forms of a vocable correspond to the given definition. In Finnish it is quite common for the occurrences of plural forms of an adjective to be nouns rather than adjectives. But of course a vocable is neither an adjective nor a noun, if all its word-occurrences of word-forms do not fit the definition normally given to adjectives and nouns in the Finnish language. Secondly it must be pointed out that generally the attempt has been made to apply to the vocables principles of classification which cannot directly be said to apply at all. As vocables have neither meaning nor any syntactical qualities, all semantic and syntactical characteristics are useless in connection with them.

The classification of words of the third type, that is vocables, is in Finnish and in most known languages, quite simple. The two self-evident groups, the "vocable classes" are inflected and uninflected vocables, in other words vocables which possess many word forms and those which have only one form. The former possesses in Finnish two sub-groups, namely the nouns and the verbs: the inflection system provides the reason for the division into these two sub-groups. Naturally there are languages in which one must create many sub-groups, but hardly any where the number of vocable classes is greater. As I have already pointed out, we cannot carry out any successful division of vocables on the basis of a semantic or syntactical point of view, for semantics and syntax are not readily applicable to vocables. Indeed, they are classes constructed on the basis of inflection, and they have no meaning or syntactical characteristics. In order to be able to classify vocables from, let us say, for example, a semological point of view, one must demand that all the word-forms and occurrences of word-forms of the vocables in question shall possess certain general semological qualities (or syntactical qualities). To enable, for example, such an extraordinarily obvious substantive (according to school grammars) as poika 3t ("a boy") to be considered to belong, on account of some semological quality, to a different vocable class from the similarly obvious adjective(according to school grammars) sininen 3t (''blue''), it would be necessary for all the forms occurring for the word poika 3t in a certain semological position to differ from the forms occurring for the word sininen 3t. The semological quality in question could, for instance, be that the substantive should always indicate things belonging to a certain class. which is at the same time stated the adjective on the other hand indicating merely: the class (itself), that is to say? a certain quality. In every instance in question the above-mentioned requirements are not fulfilled unless there is an example in which all the certain word-occurrences of the forms of poika 3t are, from a semological point of view precisely similar to those of sininen 3t and vice versa. So, for example we have in the sentence Kalakukko on poikaa ("'The fish-pie is excellent”) a proof of the case in which a word occurrence of a word, namely a word occurrence of the partitive form of poika 3t is semologically of the same type as in the sentence Kangas on sinistä (”The material is blue"), i. e. a word occurrence of the partitive form of sininen 3t. Examples of fundamentally the same type may be almost endlessly provided. To be able to put poika 3t and sininen 3t into different word-classes, however, it would be necessary that examples of this type did not exist. It may, of course, be admitted that the word occurrences of the forms of poika 3t differ semologically in most cases from the word occurrences of the forms of sininen 3t, but this fact gives us no justification for putting, the vocables poika 3t and sininen 3t into different word classes. The classes poika 3t and sininen 3t etc. do not belong together semologically and syntactically; they contain forms with word occurrences which are substantive – we can indicate this by writing poika1 3t, sininen1 3t— and adjectives in the above-mentioned sense—we can indicate this by writing poika2 3t and sininen2 3t. The traditional classification of the vocables is thus impossible. Every classification based on semological or syntactical points of view is concerned with words i. e. words of the first type certain real phenomena not with vocables. Indeed, the throwing of light upon the systematisation of the contents of a language's vocables is an interesting theme for scholarship. It is precisely on this point that great differences between different languages exist. And as we know well,—and as has been demonstrated in a remarkably clear fashion by Prof. K. G. Ljunggren, amongst others — fluctuations within the vocable can happen in the history of a language. The semologically or syntactically distinct group may grow or diminish.

The question now remains as to whether the word classification could not be built upon the forms of the words, that is words of the second type. Yet it is not easy to demonstrate how common it is for the different occurrences of the word-form to be so different semologically or syntactically that the possibility of their belonging to a mutual semological class does not exist. We may consider, for example, the word group hyvin hyvin asein ("with very good weapons"), where two occurrences of hyvin 2t possess very different meanings. We have a good example also in the series: paikalla paikalla paikalla (paikattava paikka) "(a place that must be mended) at the place at once with a patch”. From this example a similar conclusion to that arrived at with regard to the former case must be drawn, i. e., that the classification of word-forms according to semological. or syntactical principle cannot he successful.

If the doctrine of word classes were to pay regard to the facts mentioned, difficulties which cannot be surmounted would disappear. As to what word-classes (i. e. classes of words of the first type) would then exist is another question. In different languages the words must be classified in somewhat different ways. Thus the classification into partitive and non-partitive words in Finnish grammar, for example, is of much greater importance than in Swedish grammar. But on this question I cannot here enter.