Stress and Information

LEE S. HULTZEN S EXPLANATION of English stress through analogy with information theoryI comes at a time when most American linguists are busy with a different kind of horticulture, cross-pollinating stresses, or stresses and junctures, in quest of hybrids that will give a phonemic shape to syntax. It goes without saying that there are grammatical uses of stress. A simple example is the phrase steel bdr, in which the arrangement of stress is like that of green cdat, and suggests an underlying kinship that the conventional classification of steel and green obscures. But are the uses systematic? If one stress acting independently marks the information-bearing part of an utterance, it is better regarded as a morpheme than as a phoneme. On the other hand, if stress is systematically grammatical there should be some obvious key to the system. A number of keys have been offered, fitting a variety of locks. Whether the doors that they open lead to blank walls, small closets, or all outdoors is the theme of what follows. And along the way we shall try Hultzen's passkey. Hultzen criticizes Haugen's assumption 'that the stress pattern secondaryprimary applies to most adjective-noun constructions,' a notion that is 'widely accepted and manifestly false.'2 Hultzen's examples are pertinent, but the reader confirmed in the opposite point of view is apt to interpret them with the aid of the deus ex machina of the 'shift morpheme'-they are a bit too suggestive of contrastive stress. So I shall add a few where there is no suggestion of outright contrast, and where the stress on the adjective (optional, but highly probable) seems to be due to the relative pressures of information: