At a meeting of the Academic Mother Tongue Society (Akadeemiline Emakeele Selts, henceforth AES) on November 12, 1933, Ernst Nurm presented a paper titled "Are the agent and action nouns forms of the verb?" (Kas tegijanimi ja teonimi on pöördsõna vormid?).1) The paper was then published in the journal of the AES, Eesti Keel 'The Estonian Language', in its final issue of the year 1933. What follows is Nurm's argument in an abridged form:
In the early days when Estonian grammars were written in German, the agent noun (with the suffix -ja) and the action noun (with the suffix -mine) were included into the forms of the verb. They were regarded erroneously as participles. [...] All the Estonian-language grammars and exercise books have also treated these words as forms of the verb. The only differences, if any, are the names with which the forms in question are referred to in various grammars. [...] If we consider the matter more closely, however, then we will come to the conclusion that neither word can be regarded as a verb form: they should be regarded as nouns derived from the verb stem. [...] And indeed, do words in -ja differ from those in -ur in principle? Both indicate agents. Though the latter tend to refer to professions (e.g., lendur 'airman', valvur 'watchman'), words in -ja, too, often indicates persons who do something as profession (e.g., õpetaja 'teacher', kirjutaja 'writer', habemeajaja 'barber', juukselõikaja 'hairdresser', kuduja 'weaver', juurdelõikaja 'cutter [at the tailor's]', etc.). If the ending -ur is generally considered to be a suffix, then -ja should also belong to the same category. [...] As to the action noun, a similar conclusion is reached by an analogous reasoning. First of all, there is no difference in principle (and often in meaning, either) between words in -mine and those in -us: cf. kasvatamine 'upbringing' — kasvatus 'upbringing', kirjastamine 'publishing' — kirjastus 'publishing house', armastamine 'loving' — armastus 'love', ehitamine 'construction' — ehitus 'construction', and so on. [...] That the words in -mine are more of a noun than a verb become evident from the fact that the noun attribute occurring with them are put into the genitive case [...]: compare loen raamatuid 'I read books' with raamatute lugemine 'the reading of books' and raamatute lugeja 'the reader of books'; ta õpib ladina keelt 'she is studying the Latin language' with ladina keele õppimine 'study of the Latin language' and ladina keele õppija 'learner of the Latin language'. Let us then exclude, from among the verb forms, the action noun as well. [...] (Eesti Keel, 1933, 161-164)
The story is, in a sense, simple and clear: the agent and action nouns in Estonian cannot be regarded as verb forms in the strict sense of the term nor are they quite like typical nouns, either. They share features both with the verb and the noun. In terms of syntactic, morphological and semantic nature, the agent and action nouns are somewhere between the verb and the noun. Very few would disagree on this point. However, opinions can and do differ as to how this fact is to be stated in grammars, especially when we start to classify them in terms of parts of speech: are they verb forms or deverbal nouns? Until the early 1930's they had been regarded as belonging to the verb paradigm. Nurm didn't agree with this tradition and suggested that they be excluded from the verb paradigm. The AES meeting gave full support to Nurm's view of the agent and actions nouns: it was decided that they be referred to in those sections of Estonian grammar books which deal with deverbal nouns.
As is evident from the minutes of the AES meeting which appeared in the first issue of the year 1934 of Eesti Keel, the discussion was centered upon the practical significance the problem has for the writing and teaching of the normative grammar rather than the way to describe the phenomenon as objectively as possible. As one of the participants in the AES discussion put it, "nothing substantial was involved" in the transfer of the agent and action nouns from the verb paradigm to the category of derived noun, since "the question was simply where they were to be mentioned in Estonian grammar books." Nurm was eager to enumerate practical merits of his proposal with respect to the traditional view:
Now you may ask what is the practical advantage of excluding the agent and action nouns from among the verb forms. For one thing, it will reduce the number of the rather-too-many verb forms by two. It will then simplify the study of syntax to a certain degree. In many cases one makes a mistake when analyzing sentences and studying government just because the agent and action nouns are regarded as verb forms when they should always be treated syntactically as nouns. For instance, an expansion of the agent or action noun can only be an attribute. If, however, they were regarded as verb forms, their expansions would have to be regarded as objects or adverbials (as is often done by students). Similarly, if the agent noun were to be treated as a verb form, it would be difficult to explain why such expressions are incorrect as rasket tööd tegija 'doer of heavy work', mitu keelt tundja 'knower of many languages', ta on õigust armastaja inimene 'he is a lover of justice', etc. If, however, they are regarded as nouns, the reason of their incorrect-ness is quite clear. [...] As is evident from what has been said above, there is no weighty argument for treating the agent and action nouns as verb forms. There are, on the contrary, a plenty of convincing circumstances that compel us to give up the tradition and leave the two forms out of the verb paradigm as unnecessary and, every so often, misleading as well. [...] (Eesti Keel 1933, 164-165)
For what reason did the Estonian linguists of those days have to think that agent noun constructions with a partitive object such as rasket tööd tegija 'doer of heavy work', mitu keelt tundja 'knower of many languages', õigust armastaja inimene 'lover of justice', etc. were "incorrect"? It is very interesting to note in passing that Nurm was so precise that he found the "grammatical error" in question even in a writing of Andrus Saareste's! In any case, the occurrence of the partitive object in the agent noun construction seems to have been considerably more common those days than nowadays.
A short article that appeared in a journal in the early 1960's bears witness to the fact that the matter in question was far from being self-evident for the Estonians. On the page titled "Questions and Answers" (Küsimusi — vastuseid) in a 1961 issue of the journal Keel ja Kirjandus 'Language and Literature,' a "reader" asked whether it was "grammatically correct" to say abiks kunstiajalugu õppijaile 'as an aid for learners of art history'.
The anonymous authority who answered to this question based his argument on the 1933 AES decision. The essence of his answer is as follows:
Since the agent and action nouns are nouns by nature, it is not correct to say lapsi armastaja 'lover of children', muinasjuttu rääkija 'teller of a fairy tale', raamatut lugeja 'reader of a book'. Similarly, expressions like ühiskeelt kõneleja 'speaker of the standard language', füüsilist tööd tegija 'doer of physical work', filosoofiat õppija 'learner of philosophy' are not proper for the language. Instead one should say ühiskeele kõneleja, füüsilise töö tegija, filosoofia õppija. Consequently, kunstiajalugu õppija 'learner of cultural history' is not grammatically correct, either, and one should employ a genitive attribute instead: kunstiajaloo õppija. (Keel ja Kirjandus 1961, 441)
Let us recall, however, that the decision by the AES to put the agent noun into the category of noun was not made because it was a noun in the same way as the nouns such as kivi 'stone' and maa 'earth' are, but because people agreed that it was better, from a practical point of view, to regard the agent noun a noun than to regard it as a verb form. It is therefore unlikely that the AES decision was intended as a declaration which was to prescribe the change in the syntactic nature of the Estonian agent noun itself. Rather, it was simply a question of where it would be the most appropriate to mention the agent noun in Estonian grammar.
Reading the article in Keel ja Kirjandus, however, one gets the impression that its author wants to interpret the 1933 AES decision in a different way. Namely, he seems to be claiming that once it is transferred from the verb paradigm to a paragraph of the grammar dealing with nouns, it is not proper for the agent noun to share any features with the verb.
A possible explanation for what brought about this rather unusual shift in the logic of normative prescription may be found in Saareste's words at the 1933 AES meeting: "The question is also of practical importance. It is peculiar to the Finno-Ugric languages that no clear distinction is made between the noun and the verb" (Eesti Keel 1934, 31). For some reason, I suspect, Estonian linguists of those days thought that the "lack of clear distinction between the noun and the verb" was not to the advantage of the Estonian language. It was then decided to make the distinction in question a bit clearer by means of normative prescription. What was at the center of attention was the occurrence of the partitive object in the agent noun construction. There was probably no similar problem with the action noun: the partitive object does not usually occur with it, which is just what is expected of a noun. So they created a norm: the agent noun is a noun and, as a noun, it must take a genitive attribute instead of a partitive object.
It would be very interesting to investigate what motivated Estonian linguists' attempt to get rid of this "peculiar feature of the Finno-Ugric languages." If the analysis based on the traditional notion of parts of speech didn't work quite well and produced some dissonance in the treatment of the Estonian agent noun, it was, as a present-day linguist would have concluded, the traditional notion of parts of speech that was to blame. Apparently Estonian linguists thought the other way around, and tried to deform the Estonian agent noun so that it might conform to the traditional notion of parts of speech.
Half a century later, it seems that this "peculiar feature of the Finno-Ugric languages" has turned out to be considerably more tenacious than Estonian linguists thought in the 1930's. Half a century is quite a long time, and yet the Standard Estonian has not succeeded in getting rid of expressions such as jaapani keelt õppija 'learner of the Japanese language'. At the same time the environment surrounding the normative as well as the descriptive grammar of the Estonian language has gone through a fundamental change. The agent noun construction with a partitive object is no longer considered incorrect. Consult, for example, the new grammar of Standard Estonian, the publication of which is long overdue.
According to Mati Erelt, who heads the grammar project, the phenomenon in question is to be described in terms of the process which he names syntactic substantivization of participle constructions (partitsiiptarindi süntaktiline substantivatsioon):
Like an ordinary adjective, any participle functioning as an adjective can be substantivized. In the course of syntactic substantivization, the verbal-adjectival meaning of the participle is replaced by the meaning characteristic of the category of noun; in other words, the construction no longer denotes the process itself, but the person or thing that is involved in the process. The individual lexical component in its meaning, however, remains unchanged. In a similar way, the syntactic function of the participle construction also changes in the sentence, though its inner structure remains the same [...]. In syntactic substantiviza-tion, the markers of the past participles -nud and -tud get shortened and become -nu and -tu, respectively, whereas the markers of the present participles -v and -tav remain as such. The only exception is the marker -v of the present active participle: it is replaced by the suffix -ja when the subject is animate. (Erelt, 1987, 61)
Following Erelt we can say that there are three types of verbal nouns in Estonian:
Why, then, is the behavior of the agent noun rather peculiar in this particular syntactic environment? As is well known, there is no such problem with the Finnish agent noun, whose behavior with respect to the object is exactly the same as that of the action noun in the standard language: the genitive attribute alone is allowed in the agent noun construction.
The peculiarity of the Estonian agent noun can be explained by assuming that the Estonian agent noun corresponds not only to the Finnish agent noun in -ja, but also to the Finnish participle in -va with an animate subject. This amounts to say that the Estonian agent noun õppija (derived from the verb õppima 'to study, learn') corresponds both to the participle opiskeleva (present active participle of the verb opiskella 'to study') and the agent noun opiskelija (derived from the verb opiskella) in Finnish. It would mean approximately that the Estonian construction jaapani keelt õppija 'learner of the Japanese language' corresponds to the Finnish expression japanin kieltä opiskeleva 'one who studies the Japanese language' and the Estonian construction jaapani keele õppija 'learner of the Japanese language' to the Finnish expression japanin kielen opiskelija 'student of the Japanese language'.
Like the Finnish agent noun, the Estonian agent noun has been in constant competition with the participle in -v. The Finnish agent noun has had to give way to the participle in -va and, in so doing, has had to lose its ability to take an object and moved away from the verb in the direction of a full noun. The Estonian agent noun, on the other hand, has been resistant to a similar tendency and, rather, acquired a new territory from the hands of the participle in -v: if the subject is animate, the substantivization of the participle in -v takes place, unlike the Finnish language, by means of the agent noun instead of the participle itself.
When collecting examples of agent and action noun constructions from Estonian newspapers, I noticed a very interesting difference in the inner structures of the two constructions. Namely, the occurence of the da-infinitive in the agent noun construction is very common as in karjääri teha soovija 'one who wishes to advance one's career', whereas the same phenomenon is practically absent in the action noun construction: though the action noun construction is considerably more frequent than the agent noun construction, I have not yet encountered a case in which the da-infinitive occurs in front of the action noun. This is one of the indications of the greater viability the Estonian agent noun possesses compared with the Finnish agent noun.
*) This is an English translation of the paper which I presented orally in Estonian on August 31, 1990, in Debrecen, Hungary, at the Seventh International Congress of Finno-Ugrists (CIFU-7). The Estonian version was published in the proceedings of the congress: Congressus Septimus Internationalis Fenno-Ugristarum, 3B: Linguistica, 1990, 321-325. [Back to Main Text ]
1) The Estonian terms tegijanimi 'nomen agentis' and teonimi 'nomen actionis' are rendered into English as 'agent noun' and 'action noun' in the present paper. [Back to Main Text ]