The notion of ‘subject’ as a grammatical category first appeared in Western Europe about the 12th to 13th century by the name of ‘suppositum’ and was theoretically elaborated on by the Medieval grammarians called ‘modistae’. Before that, however, it was quite unknown, surprising indeed, in any grammatical tradition of the world ; neither Apollonios, nor Priscianus, nor Panini, nor Arab grammarians knew such a category.
The present lecture aims to elucidate the reason why the notion of subject did appear in Medieval Europe but not in other parts of the world, and the linguistic motivations for its appearance in Europe or its absence elsewhere. I shall discuss, especially, the theory of karaka of Panini, who completely dispenses with syntactic relations such as subject and object, the Stoic theory of kategorema, which foreshadows the later subject-predicate notion, the theory of mubtada'/khabar (=Topic/Comment) and of ‘amil/ma ’ mul (=regens/rectum) of the Arab grammarians after the 8th century, and finally the theory of suppositum/appositum (=subject/predicate) and principium/terminus (=the relation of ‘subject of ’/ ‘object of’) of the European modista.
The appearance of subject in the West European grammar, it will be pointed out, was linguistically conditioned by the development of the strict SVO word-order combined with the characteristic morphological attrition, especially the loss of nominal cases and of verbal person markings, which seems to have been brought about by the creole-like processes resulting from the bi- or multilingual situations among the Germanic and Romance speaking peoples in close contact during the Medieval age. This can most typically be seen in the case of Middle English. Thus, thesubject as a syntactic category is really a historical product in a quite limited linguistic area which comprises those languages once called SAE (= “Standard Average European ”) by B. L. Whorf. They share a typologically unique feature known as “dummy subject” and thus manifesting themseleves as so-called “non-pro-drop” languages.
In short, the subject of SAE has resulted from the coalescence of three quite different linguistic functions into a purely syntactic category, namely, 1) the discourse topic, which was the very starting point of the Medieval notion of suppositum, 2) the morphological case-marking (i. e. nominative vs. accusative), and 3) the semantic agent (or rather the “primary role”) of a verb. Usually, these functions are separately grammaticalized in other languages, e. g. as mubtada' and fail in Arabic, as wa and ga in Japanese, as “focus” and case in Philippine languages, as word-order, case and verbal endings in Old Indo-European languages.
The subject, in conclusion, cannot be part of the theory of syntax as a universal category. It is quite a complex and heterogeneous concept in its origin, and manifests as surface syntactic phenomena only in a very limited number of -languages. Therefore, any syntactic theory based exclusively on the observations of such languages needs to be thoroughly reexamined, if it claims to be universal in any sense. Rather, we must reappraise the old grammatical traditions radically different from that of West European or “Aristotelo-Cartesian” school in order to found a theory of “Universal Grammar” in the true sense.
(This is a revised version of the Presidential Lecture given at the 102nd General Meeting of the Linguistic Society of Japan held at Tokyo University of Foreign Studies, June 6th, 1991).

View full abstract

This paper argues against the traditional view that Japanese pronouns differ from English pronouns in that they can never be bound by an operator. Rather, it is shown that pronouns cross-linguistically have the property of ' Anti-Minimality, 'by which is meant that a pronoun can be operator-bound only if another operator intervenes between the pronoun and its binding operator. Thus, just as pronouns stand opposed to anaphors with respect to A-binding, so they also stand opposed to variables with respect to O-binding; both anaphors and variables have the property of Minimality for the relevant type of binding. Data from English, Chinese and Japanese are discussed, and it is argued that discrepancies in the distribution of operator-bound pronouns derive from the existence/nonexistence of the pertaining Anti-Minimality inducers in these languages, which in turn reduces to the parametric variations in their configurational structure. Thus English differs from Chinese and Japanese in that only the former has AGR, while Japanese differs from the other two in that it crucially lacks a subject. Hence the strictly limited occurrence of operator-bound pronouns in this language. Since this study shows that the concept of Minimality comes into play for binding as well as for government, it provides further motivation for the reduction of the ECP to the Binding theory.

View full abstract

The aim of this paper is' to formulate a rule which assigns the mora phoneme /Q/ to English loan words in Japanese. To achieve this aim, I will ' assume that morae should be classified into three types with respect to "weight" : normal mora consisting of CV is regarded as weight 1, super mora consisting of CVC or CV as weight 2, and ultra-super mora consisting of CVC as weight 3. It will be found from this assumption that /Q/ is inserted between two morae when the difference of the weight of the two morae is 1. In other words, when the weight difference is either the same or more than' 1, /Q/ is not inserted. I will formulate this finding as /Q/-Insertion Rule, and attempt to explain in terms of this rule why some words like hit and ' tap have /Q/ in their loan forms while others like carp and bike do not. In addition to /Q/-Insertion Rule, an extrametrical rule is also proposed which handles words like dress and tough exceptional to the insertion rule.

View full abstract

Progress in Tocharian studies is very slow; there still remain a large number of problems which resist unanimous agreement. This is due mainly to a chain of complex changes which took place in the prehistory of both of the Tocharian languages. Recent developments in historical phonology have, however, given us a much clearer idea of the evolution of the Indo-European sound system in Tocharian. On the basis of the results thus achieved in phonology this paper attempts to offer a historical explanation for the morphological issues involving the following three sets of forms: (1)TA wal, TB walo 'king', TB kamem'(they) came', (2)TA rasna'(he) would stretch', and (3) TA nom 'name', TB nem, nemantse.

View full abstract