Archibald A. Hill

Word Stress and the Suffix -ic

earlier, has attracted notice at least since 1791. In that year John Walker, writing in his Principles of English Pronunciation, remarked that, &dquo;... in words ending in ic, the accent immediately precedes the termination .... The only exceptions to this rule are, arsenic, choleric, ephemeric, turmeric, empiric, rhetoric, bishopric ... lunatic, arithmetic, splenetic, heretic, politic, and perhaps phlegmatic.... &dquo;~ The modern discussion of the suffix and its stress-patterns begins with the NED, under the entry -ic. In this entry, presumably written by James A. H. Murray, it is stated, &dquo;Words in -ic from Gr. or L. have the stress

Juncture and Syllable Division in Latin

There have been a number of successful attempts, such as Kents,i to explain variations in scansion or in the development of Vulgar Latin forms in terms of variations in syllable division. It is also rather commonly accepted that Latin had some means of marking word boundaries-that is to say, had a juncture phoneme. There has not been (to my knowledge at least) any attempt to put syllable division and juncture together in a systematic fashion, to see what each can tell us about the other and what both can tell us about the phonemic structure of the language. Such a synthesis is the task here attempted. It is the contention of this paper that when the two are put together, three situation types emerge: normal syllable division, normal syllable division accompanied by juncture, and syllable division of a type that requires juncture. In syllable division we have evidence from scansion, from certain inscriptions where syllables are divided by points, and from direct statements by the grammarians; these several kinds of evidence establish the following facts. (1) A syllable is long if it contains a diphthong or a long vowel, or ends in one or more consonants. (2) The sequence /VCV/2 is normally divided /V-CV/, not /VC-V/, so that the prior syllable is short. (3) The sequence /VCCV/, at least with non-initial clusters, is normally divided between the consonants, so that the prior syllable is long. (4) If there are more than two consonants in medial position, there is likely to be a division somewhere within the cluster, but scansion evidence will not tell us where, since scansion does not distinguish between syllables which end in one and those which end in more than one consonant.

Post-nominal modifiers: Transformations and phonology

Abstract In my Introduction to Linguistic Structures, I reported the results of the first of a series of tests on sentences with restrictive and non-restrictive elements as modifiers of a noun subject. 1 I have since then carried these tests somewhat further, and can now report more far-reaching results. The tests consisted in pronouncing possibly ambiguous sentences with various types of pitch and terminal patterns, to see which, if any, of these patterns resulted in convergence of identification by groups of native speakers. The results, simply stated, are that only two pitch-terminal patterns are unambiguous phonological signals, and that of three types of sentence where a following element may be functionally restrictive or non-restrictive, there is one in which even the otherwise clear phonological signals fail, leaving the sentence completely ambiguous.

SOME THOUGHTS ON SEGMENTATION OF LEXICAL MEANING

At a meeting of the Linguistic Society in 1947, I heard the late Edgar Sturtevant declare flatly that there was not and never would be a scientific semantics. As to whether there was a scientific semantics in existence then, I think that all linguists whose memory goes that far back would agreethere were only relatively superficial attempts like the Ogden and Richards Meaning of Meaning, and the somewhat overblown pronouncements of the Korzybskian school. But as a prophecy, it now, seems certain that Sturtevant was wrong. Nearly all linguists would agree that there have been a number of important contributions to the study of meaning, and that they are continuing. The difficulties in describing the meaning of words are of course still with us, though our successes have been considerable. The basic difficulty in 1947 appeared to be double. One was that there was no sharp dividing line between classes .of objects designated by different words, such as desk and table, and growing out of this fact, the further fact that the only ways of defining desk and table as different words was to rely on the absence of similarity in sound, or to say that they meant different things. The first method was unsatisfactorily mechanical and even arbitrary, since most linguists would have defined g o and went as forms of the same word in spite of lack of phonological similarity, and the second method was felt to be unsatisfactory because of circularity-that is, one derived meaning from the recognition of words, and recognized words by their meaning. I can trace my own thinking about such problems as these through a number of steps. The first is the distinction in kinds of meaning which I owe to Henry Lee Smith and George Trager. These scholars used the concept of “differential meaning,” as something quite apart from referential meaning, and set up use of this kind of meaning as a way of escaping from the circularity of which I have just spoken. That is, mat and gnat are identified as different words by the sounds which occur in initial position, and the different identity of the two words proves that the initial sounds are different. The dispute over whether the use of “differential meaning” was really legitimate simmered for years, and there were bold and persistent attempts by both Bloch and Harris to describe the sounds of English by the use of distribution alone, without any reference to meaning at all, either differential or referential. The attempts were certainly necessary, since no one could know that they would fail until they had been thoroughly tried, but it seems to me now that the reputation American linguistics earned for being anti-semantic was largely due to the fact that we did not try other avenues of identifying meaningful forms until the slow process of trying to use distribution alone had been shown to be a failure. For me, at any rate, a second stage in arriving at a position on identification of forms comes from a rather curious place, curious, however, only because it

Some speculations on tempo in speech

This article is contributed in memory of Claude Wise. Its subject is the focus of a conversation between Wise and the author on the occasion of one of this distinguished teachers visits to Austin, Texas, during his green and vigorous years of retirement.