sis is occasionally used. The latter is meant to indicate a breaking off in the formation of a sentence, or even of a combination of sentences, an interruption that is caused by rhetorical motives, or by an overwhelming emotion. Examples are afforded by the three last mentioned quotations. But it is left undetermined whether aposiopesis is only a subdivision of ellipsis or of brachylogy, or whether it is to be coordinated with them. Nor does there seem to be any line of demarcation drawn between the former and the two latter.

It is perhaps superfluous to point out that the inward discourse may appear in its outward form in a more or less modified shape. It is not only the morphems which present such a close resemblance to the inward discourse as Mr. Jingle's language that should be termed reflections of inward discourse. In our opinion, any sentence with incomplete resolution may be classed as such a reflection, consequently also the categories of sentences we are now going to consider.

It is a matter of general observation that a morphem may stand in relation to more than one morphem, or, as PAUL1 puts it, there occurs 'einmalige Setzung mit mehrfacher Beziehung'. Such apokoinou-constructions are often met with, especially in paratactical combinations of sentences, e. g. Stones have been known to move and trees to speak (Sh., Macbeth, Act. III, sc. 4), Many have and others must sit there (Sh., Richard II, Act V, sc. 5). The same construction is also very frequently found in questions and answers, for even the dialogue »muss als etwas kontinuierlich Zusammenhängendes betrachtet werden«. If we ask Who is singing? and the answer is Mary, or if we say She sang beautifully and some one asks Whor, then the words Mary and who have the nature of sentences through their close connection with the preceding morphems, which afford the necessary material for their reconstruction into sentences.

For both these cases grammar generally employs the term ellipsis, sometimes, however, also the term brachylogy. In both these cases too, ellipsis is tantamount to incomplete resolution. It must be noted, however, that it is here possible to reconstruct with certainty the corresponding full expressions. For we are here conscious of which elements are left unexpressed, and this in its turn is due to the circumstance that they have been ac

1 Cf. PAUL, op. cit., p. 290.

2 Cf. Ibid., p. 291.

tually embodied in language, though this is performed in connection with other morphems immediately preceding or following.

Of a special kind is the incomplete resolution originating when the speaker does not master a language. Here belong, not only the cases when a person is expressing himself in a foreign language he does not command, but also the first stages of children's speech before the faculty of complete resolution has been acquired. As illustrative instances of baby-talk we may take from STEINTHAL1 the following German examples: »Mama - Baba = Ich will bei der Mutter schlafen, Papa - Hut der Vater hat einen Hut auf, Dat Huhn = der Soldat sitzt zu Pferde. »

=

Some, perhaps, will be disinclined to recognize such morphems as sentences. But what are they then, and what becomes then of the principle that all linguistic utterance consists in sentence-formation? They are indisputably sentences, as is pointed out by WUNDT and GOMPERZ3. They cannot therefore be designated as futile attempts at framing sentences, but instead as examples of abortive efforts to resolve completely a thought into its elements. Consistency should therefore demand that they should be classed by grammar as elliptical.

We know that in written language morphems may occur which in themselves do not signify a sentence, but get such signification owing to the circumstances in which they appear. This is especially the case with inscriptions and the language on tickets, signboards, etc. If, for instance, we read on a building Haymarket Theatre, we infer from the location that the inscription. is equivalent to the sentence 'This house is the Haymarket Theatre', or the like. DELBRÜCK says that in such cases it is dubious >>ob eine Abbreviatur eines ursprunglichen Satzes vorliegt. But they ought to be interpreted as representing an intentional combination of written language with information inferred from the locality. What has not been expressed through the inscription is sufficiently expressed by the location to let the reader reconstruct the total representation intended, with the same result as if it had been completely resolved into separate morphems. When such sentences are termed elliptical, ellipsis is again tantamount to incomplete resolution.

1 Cf. STEINTHAL, Abriss der Sprachwissenschaft, i. 399 sqq.

2 Cf. WUNDT, op. cit., p. 232.

3 Cf. GOMPERZ, op. cit., p. 52.

4 Cf. DELBRÜCK, op. cit., p. 142

The second category of sentences classed as elliptical that we have distinguished, is the one, where the sentences offering an incomplete resolution, may be considered as survivals of older types of sentences.

Such exclamative sentences as Fire!, Murder!, Thieves!, etc. are no doubt incomplety resolved. It is only the predominating element of the thought that has been expressed, while the nature of a sentence is lent to the morphems by their emotional intonation. They must be distinguished from sentence-equivalents met with in dialogues or in narrative style, where it is context and attending circumstances that give them the character of sentences. They must not, like the latter, be looked upon as an incomplete application of the modes of resolution current in Indo-European languages. These rudimentary sentences ought rather to be considered to bear record of a bygone past. They are venerable survivals from the childhood of mankind, when the origin of language was no doubt to a considerable degree conditioned by such exclamative expressions. When they are designated as elliptical', it is no doubt because they are incompletely resolved. But as ellipsis pretends to be an historical means of interpretation, its use here conveys the false idea that the morphems have originated through incomplete application of the predicative mode of moulding sentences.

The third category of sentences designated as elliptical, comprised such which, comparatively speaking, were accurately split up, but deviated from the common type.

Wundt recognizes two different principal modes of splitting up thoughts into sentences, namely, the predicative and the attributive type. By predicative sentences he means such as have been split up into a subject and a verbal predicate, e. g. G. Der Blitz leuchtet, Der Donner rollt, Die Rose ist roth2. By attributive sentences he seems to designate such as have been split up into a substantive and its attribute, and which consequently are lacking in a verbal predicate, e. g. G. Welche Freude!, Welch eine Wendung durch Gottes Fügung!3 The parsing of the latter sentence is given by WUNDT in the following way: »Ein Subject hat ein solcher attributiver Satz offenbar ebenso gut wie ein prädikativer: die Vor

1 Cf. J. Grimm, Deuts. Gr., iii., p. 306, iv., p. 131 (foot-note).

2 Cf. WUNDT, op. cit., p. 317.

s Cf. Ibid., p. 264.

Cf. Ibid., p. 266.

stellung, die hier die Grundlage des Ausdrucks bildet, ist das worauf sich das Attribut bezieht. Statt eines Prädicates hat er aber ein Attribut, d. h. er enthält keinen zweiten Begriff, der von dem ersten ausgesagt werden soll, sondern statt dessen eine nähere Bestimmung, die zu jenem hinzugefügt wird. In diesem Sinne ist welch eine Wendung das Subject, durch Gottes Fügung das Attribut des Satzes». He adds: >>Die gleichen Verhältnisse ergeben sich, wo immer wir Gefühlssätze mit rein nominalen Ausdrucksformen in ihre Bestandtheile zerlegen, mögen jene nun von einfacher oder zusammengesetzter Beschaffenheit sein. In der That ist schon der obige Satz nicht von ganz einfacher Structur. Sein Subject welch eine Wendung würde für sich allein einen vollkommen zureichenden Gefühlssatz bilden. In diesem Falle würden wir aber die Wendung als den tragenden Subjectsbegriff, das zum Ausruf verwendete und durch den Artikel mit dem Subject verbundene Pronomen welch eine als das Attribut zu betrachten haben: diese Pronominalverbindung fügt dem Subject einen stark gefühlsbetonten Hinweis hinzu, der logisch wieder nur als Attribut, als eine dem Hauptbegriff beigegebene nähere Bestimmung, nicht als eine prädicierende Aussage über ihn aufgefasst werden kann. Die dominierende Vorstellung ist aber in diesem einfacheren Satze das Attribut welch eine, wäh rend sie in dem zusammengesetzteren das ganze logische Subject welch eine Wendung war».

The predicative form of framing thoughts in language is considered by WUNDT to be indicative of a high evolution of thought, which is in accordance with the fact that this form chiefly occurs in highly developed languages1. Again, the attributive form is principally met with in languages on a low scale of evolution, and seems therefore to have priority to the predicative type. Thus, in the Fiji language one says Heart or will-of-me instead of I will, and in other Polynesian languages the sentence I will eat the rice is expressed by The-eating-of-me-the-rice, etc.. The predicative type is the prevalent one in Indo-European languages, but when in them such exclamative sentences as the above-mentioned are met with, they ought, in WUNDT's opinion, to be looked upon as representatives of the attributive form of thought. If this interpretation is right, it follows that they must not be explained

1 Cf. WUNDT, op. cit., p. 328.

2 Cf. G. J. Romanes, Mental Evolution in Man (Origin of Human Faculty) London, 1888, p. 318.

as elliptical, as is generally done. For this would imply that they offer an incomplete application of the predicative mode of resolving a total representation into its elements.

Such sentences, too, as G. Lumpenhunde die Reiter (Göthe), Lat. omnia præclara rara, are by WUNDT reckoned to the attributive type. He maintains that the copula has exclusively a formal function, the one namely, of transforming an originally attributive expression into a predicative one, although he admits that the sentences 'dem Gehalte nach' remain attributive1. But DELBRÜCK points out, in our opinion justly, that the logical relation between the parts of the sentences is in nowise changed by adding the copula. The only difference, he says, lies in the fact that the relation in one case is expressed by a special formword, in the other by stress and intonation. He adds: »Ich würde es unter diesen Umständen für praktisch halten, den Ausdruck attributiv dem Falle vorzubehalten, wo ein Adjectivum einen Theil der Benennung eines Gegenstandes bildet, z. B. der grüne Baum, dagegen den Gebrauch in grün der Baum nach wie vor prädikativ zu nennen. Die attributiven Sätze Wundt's würden vielleicht passend als nominale Sätze bezeichnet werden. It should be observed, however, that both WUNDT and DELBRÜCK concur with LUGEBIL in his opinion that sentences of this stamp must not be interpreted as elliptical because of their want of the copula, but instead as survivals of an ancient type of sentence lacking in a verb. For, as they justly affirm, the formal function of the copula has gradually been developed out of a 'Verbum mit anschaulicher Bedeutung'. Consequently we have to do with an additional example of a wrong use of ellipsis as a means of interpretation. And this wrong use arises once more from the aprioristic view that the predicative form of thought is the original and only one, which is tantamount to the opinion that noun and verb are indispensable constituents of any complete sentence.

It seems as if WUNDT were inclined to recognize as attributive only such sentences, consisting of a morphem determined by an attribute, which are of emotional nature ('Gefühlssätze'). But it is questionable whether consistency should not require us also to

1 Cf. WUNDT, op. cit., p. 260.

* Cf. B. DELBRÜCK, Grundfragen der Sprachforschung, Strassburg, 1901, p. 149. Cf. LUGEBIL in Archiv für slav. Philologie, herausgegeben von Jagič, Berlin, 1888, vol. viii., p. 36.

4 Cf. B. DELBRÜCK, op. cit., p. 148.