The notion of a word is subject to various ambiguities. The text of the Hebrew Bible (Elliger and Rudolph 1977) provides us with several conflicting notions of word. The Biblical text consists of two main layers, and the written word - in the sense of letters surrounded by blank space - differs in each layer.
The earlier layer contains a consonantal text, devoid of almost all indications of vowelling and punctuation. I will show that the word in this consonantal layer corresponds to a potential prosodic word, that is, a unit that could be an independent word for purposes of phrasing, whether or not it actually functions as such in any particular context. Since such words are not necessarily prosodic words in every context, I will call the word in this layer an orthographic word. Ordinary written Hebrew makes similar word divisions.
To the consonantal text were later added various diacritic marks indicating vowels, some allophonic consonantal distinctions, and an elaborate system of "accents" that indicates position of stress, division into verses, and a highly articulated prosodic parse of each verse. Orthographic words of the consonantal text could be joined together by hyphens to create a larger unit, the prosodic word. I will show that the principles for forming prosodic words Ð rules of cliticization Ð are quite complex, and interact in intricate ways with other aspects of prosodic structure, such as the phonological phrase and the intonational phrase.
Turning to the evidence of the phonology, I will distinguish between the prosodic word and the phonological word, which is the notion of word referred to by the phonology proper (segmental processes, syllabification, stress), as opposed to the phrasing. Here I will show that there are two competing notions of phonological word that must be disentangled. Though the phonological word necessarily has some relation to the prosodic word, neither of the two characterizations of the phonological word is identical to the prosodic word. I will suggest that what must have begun as a unitary prosodic category over time began to take on morphosyntactic conditioning. In a certain class of forms, this development appears to have taken place on a word-by-word basis, producing highly variable results. Thus, a study of the word in Biblical Hebrew bears on issues of the syntax-phonology mapping in contemporary linguistic theory, as well as on the notion of levels in Lexical Phonology and Morphology.