so that English stone (noun) and stone (adjective) or German schlecht (adjective) and schlecht (adverb) may be considered varieties of each other or even the same words. The aloofness of noun from verb is more pronounced, so that act (noun) and act (verb) might be more strictly kept apart. However, there seems to be a great deal of hesitation, with a tendency to relative identification. But if a phonetic difference is found to complicate the situation, as for instance in English [haus] (noun) and [hauz] (verb), this tendency loses much of its strength. (b) A morphological difference in the technical sense of the word is of course always semanto-phonetic. Now, it may happen that one of the two morphological parts (i.e. the radical or the non-radical element) is the same, or that both are different. If there are two different radicals with identical non-radical elements we speak of two different words. For instance, hand-s and arm-s. If the same radical is expanded by two different non-radical elements, the absolute differentiation or relative identification again depends upon the relative aloofness or closeness between the differentiating parts. In English, German, and French a difference of gender, number, comparative degree, conjugational or declensional form is generally thought to be slight enough to allow of relative identification. Other differences of categorical or non-categorical meaning are generally felt to entail word-differentiation. Thus in English, love: lovely: loveliness, (to) fall: (to) fell, yellow: yellowish; or in German, können: kennen, hoch: Höhe: Hoheit, kommen: bekommen, Furcht: furchtbar: furchtsam; or in French, travailler: travailleur, etc. If both the radical and the non-radical elements are different, the relative semantic identity may occasionally be felt so strongly that relative word-identification is the result. Thus the various conjugational forms of to be might be said to be varieties of one another. As a rule, however, the feeling of differentiation will prevail. Without attempting to give any formal definitions we may sum up our discussion by pointing out such factors as may be inferred to be relevant or irrelevant to the sentence and the word. For the sentence, some articulate sound or sounds are necessary; their number or character is immaterial, provided there is one; they are determined partly by external circumstances, inasmuch as they are utilized systematic symbols grouped according to the prevailing system of sentence-structure, and partly by free choice, inasmuch as we may choose from a variety of symbols and structures and run them together to a degree suitable for the purpose; each systematic word-symbol utilized must be present at least in the form of a sound-element; the sound or sound-combination may be in a momentary sign-context with the speaker's attitude and uttered either for the purpose of mere selfexpression or simultaneously also of communication; the semanto-phonetic unit so obtained must be capable of producing a similar attitude in a suitable interpreter; or the sound or sound-combination may be in a momentary symbol-context with a specified reference and uttered for the purpose of communication; the semanto-phonetic unit so obtained must be capable of causing a suitable interpreter to make the same reference in all relevant respects. For the word, some articulate sound or sounds are required; their number and character are immaterial, provided there is one; they are predetermined by external circumstances, historical and systematic; they are systematically grouped as a unit; they are in a mnemonically predetermined symbol-context with a complex reference or consciously interrelated references likewise systematically grouped and relatively determined; this semanto-phonetic unit is in its basic features common to all the speakers of the same language, provided it is part of their vocabulary; possible complications with sign-contexts are immaterial; this unit is ready to become sentence-material without needing any further systematic structure or allowing of any decomposition into other such units; the use made of it by the sentence is immaterial, provided a phonetic element is left whose corresponding specified reference is traced to the systematic word-reference; both self-expression and communication are foreign to the word. On this basis it is possible to define a) the simple word as a word which is felt by the speaking community not to be decomposable into or synthetically resulting from minor semanto-phonetic parts; b) the morphological word as a word which is felt by the speaking community to result from a combination of two minor semanto-phonetic parts, at least one of which cannot become sentence-material without further systematic structure; c) the compound word, if primary, as a word felt by the speaking community to result from a combination of two simple or morphological words whose phonetic elements are united according to a systematic pattern and each of whose referential contexts appropriates part of the other. For the word within the sentence, some sound, sound-combination, or sound-element (nasalization, accent, timbre, etc.) is required; the character or number of these features is immaterial; they are determined partly by the systematic word, partly by the degree of absorption and adaptation in the sentence; they are in a momentary sign- or symbol context with a semantic part of the sentence, but as such they form no unit; their emotive or referential meaning as well as their phonetic part is found in some corresponding systematic word-context, or is felt by the speaking community to be traceable to some such context. Finally we may note that the word as described above is clearly distinguished from the dictionary-word.29 The dictionary, it is true, gives the main referential contexts of a word; but because its classification is founded exclusively upon the arbitrary principle of alphabetical order, its words may also include such references as are felt by the speaking community to be part of entirely separate word-contexts (e.g. ball 'round object', and ball 'entertainment of dancing'). Besides, hardly anything of the systematic phonetic grouping is indicated. To give a somewhat faithful picture of the system-word the dictionary would have to group words together in clusters showing the many-sided phonetic and semantic contexts of each word. 29 Cf. L. Weisgerber 318, 319; A. Dauzat, La Philosophie du Langage, pp. 214215 (1917). BOOK REVIEWS Menomini Texts. Pp. xiv + 607. By LEONARD BLOOMFIELD., New York: American Ethnological Society (G. E. Stechert and Co., Agents), 1928. Professor Bloomfield has given us here a series of connected Menomini texts, adequately recorded, together with English translation. Hitherto the best Menomini linguistic material has been the isolated words given by him; the words cited by Hoffman, etc., are so badly recorded as to be of little value for detailed studies. Now we have text material, ample in extent and of a high quality, upon which a descriptive (not to say comparative) grammar may be based. A bibliography on the Menomini in general is given, from which hardly anything of importance is missing, though perhaps Boas' work on their physical anthropology (in Z. f. Ethnologie, 1895) and Dixon's 'Mythology of the Central and Eastern Algonkins' (J. American Folk-Lore XXII) might well have been included. A handsome tribute is paid the late Alanson Skinner for his labors among the Menomini; and due acknowledgments are given to his own informants, etc. The volume is in general faultless in externals; personally I should have preferred roman to italic type for the Indian texts, and an indication of every five lines (or some other arrangement) would have made references to the texts easier. The only misprint in English which I have noted is 'whar' (263) for 'what.' So far as I can judge the English translation is close to the Indian texts. Occasionally it deviates slightly. So, for example, nipāt (354) is translated 'he went to bed', whereas, to judge from the analogy of several other Algonquian languages, it means 'he slept'. And occasionally passives are translated as actives. So ikō'kin (258) is rendered 'he told them'. These are minor points, perhaps not worth mentioning when reviewing so splendid a piece of work. Menomini is not an archaic Algonquian language; there are a number of secondary phonetic changes which must be borne in mind when comparing it with other Algonquian languages; but there are also variations which are inherited: among which I think the contrast between the u of uke'maw 'chief' (280) and ō of kitō'kimaminaw 'our [incl.] chief' (282) is to be counted, as shown by the evidence of Fox, Kickapoo, Sauk, and Ojibwa. On the other hand Menomini is at times decidedly useful. So, for example, ki'waskipi'w 'he is drunk' (280) backs my Fox kiwa'ckwäpyäwa as opposed to Jones' kīwe-; Algonkin proper, Kickapoo, Ojibwa, and Sauk fall in line also. More archaic than Fox A- is Menomini u- in unā'kōw 'yesterday' (Fox Anagōwe), usa'm 'too much' (356; Fox A'sāmi), usē'htawak 'they made it' (570; Fox A'ci'tōwagi), uhka'nan 'bones' (404; Fox A'kanani), etc.; similarly the i of iskō'täw 'fire' (2; Fox A'ckutäwi). The evidence of Cree and Ojibwa proves the secondary character of Fox A- in such cases. Yet the chronology of these is not the same everywhere, as is shown by Kickapoo which sometimes agrees with Cree and Ojibwa as regards the quality of the vowel, sometimes with Fox (Jones' Kickapoo texts can not be entirely relied upon for the distinctions). It should be mentioned that even at times the strongly aberrant Arapaho shows Fox A- to be unoriginal. Peoria (and Lenape) supports Bloomfield's contention that the surd stops of Menomini are more original than the sonant stops of Fox, Cree, and Ojibwa; so does Blackfoot. At times Menomini apparently is our sole guide in reconstructing the Algonquian archetypes; so, for example, the y after the ts in as unitsya'nehsitua? 'when they had a child' (76), as Bloomfield himself has pointed out. The contrast between Menomini kitäsku'ahtem 'your door' (430) and Fox ketō'ckwātämi is due to secondary changes in both Menomini and Fox; in Fox they are, I think, partly not strictly phonetic, but rather analogical; per contra compare the regular set Menomini isku'ahtem 'door' (2), Fox a'ckwätämi. Owing to the fact that rhetorical and allegro forms are very properly left (and not normalized), occasionally it is difficult for one not wholly familiar with Menomini to be quite certain as to what the original quality of a given vowel is. A list of stems with references to the texts would have been a fine help in such matters. Let us hope that we will not have to wait too long for a Menomini vocabulary and grammar. In the meanwhile let us congratulate ourselves that Bloomfield has given us a tool which every one can use for his own advantage. TRUMAN MICHELSON Comparison of Tagalog and Iloko: Dissertation for the obtention (sic) of the Doctor's degree in the Faculty of Philosophy of the University of Hamburg: presented by Cecilio Lopez of Manila: Hamburg 1928: pp. 188. This study is the work of a young Tagalog, formerly an instructor in French at the University of the Philippines and at the same time a stu |