Hebrew morphology by itself

Regarding the choice of which model most adequately represents Hebrew
morphology, it is argued that such a model must arise from Hebrew facts, that is
Hebrew morphology by itself. Specifically, this necessitates theoretical and empirical
understanding of the essence of the Hebrew root. Three models are considered:
Word-based (WB), Root-based (RB) and Root-pattern based (RPB). We show that the
nature of Semitic/Hebrew morphology is such that new lexemes (as well as nonsense
or potential words) were and still are generated directly from roots and patterns
without any need and any factual support for an intermediate word stage. The WB
model (Aronoff 1976) we conclude is inadequate for mainstream Hebrew
morphology. We then show that a Hebrew root morpheme is not a stem undergoing
vowel modification. Nor can it merge with vowels, or with concatenative affixes
(derivational or inflectional). It can only merge with a pattern morpheme: one root
morpheme with one pattern morpheme at a time. This maxim is lost in a RB model.
We propose an alternative, namely, root-pattern-based model. We show that such a
model succeeds in representing the equal, necessary, exclusive bond that holds
between the root morpheme and the pattern morpheme. Such a model is supported by
empirical data. We conclude with a brief glance at what we consider the phonological
and semantic nature of the root-pattern model.