Arabic Degree Semantics

A fundamental human cognitive skill is the ability to determine and compare the degree to which things exemplify qualitieshow tall someone is, how warm the water is, how angry the neighbor is, etc. Degree is so prominent in humans perception of their environment that it is commonly encoded as a grammatical category in human languages in the form of compara- tive and superlative forms of adjectives (taller, tallest) and words like more, less, very, enough, etc., referred to as degree modifiers. In addition to reflecting a fundamental percep- tual category, degree modifiers are a significant testing ground for two types of theories of meaning in human language: theories that interpret sentences directly as they are, and theories that posit mental rearrangements to the sentence before it is interpreted. In English and near- ly all other languages documented in this respect, degree modifiers occur directly next to the term whose amount they specify, as in Sarah drank the most coffee, where most is the degree modifier and coffee the scalar associate. While one type of theory derives the meaning of this sentence directly from the word order that occurs there, another type of theory claims that this sentence is mentally transformed into something along the lines of Sarah most drank cof- fee, and then that sentence is interpreted directly, as a comparison between Sarah and others to whom the phrase drank coffee applies. The transformational theory is motivated by indirect semantic evidence, since the transformation is not visible. But it has recently been discovered that in some dialects of modern Arabic, the coffee comparison above can visibly take rough- ly the form of the second sentence, where the degree modifier is separated from the scalar associate. The situation in Arabic allows us to evaluate the two theories in the context of a language whose surface word order mirrors what the transformational theory says is a direct representation of its meaning. This project investigates several specific constructions in Ara- bic that in better studied languages have served as testing grounds for theories of degree, while collecting new data from Syrian Arabic speakers living in Austria using techniques in semantic field methods. The Arabic pattern is exceedingly rare among the worlds languages, meaning that Arabic provides us with a unique window into the underlying architecture of the grammar of degree that is not provided by better studied languages. With this new data, we will be able to break through the confines of what can be known on the basis of previously studied languages and remove longstanding barriers to progress in the theory of meaning.

This project has sought to capitalize on a newly discovered, cross-linguistically rare, grammatical property of Arabic. In Arabic, gradable words-that can be said to hold to an extent or degree-may project their gradability to a phrase containing them, so that that phrase becomes gradable. As a consequence, comparative and superlative words ('more' and 'most', 'less' and 'least'), which have to combine with a gradable word, may in Arabic combine with a phrase containing a gradable word, resulting in phrases like-literally-'the most one who borrowed books from the library' meaning 'the person who borrowed the most books'. Such unusual locutions present an opportunity to learn more about the grammar of gradability. This project looked at four specific questions in this connection. One concerns whether projection helps clarify how the superlative word 'most' is related to its inverse counterpart 'least'. In Arabic, projection can 'split' the word for 'least' into the component parts 'most' and negation, showing that inversion is the effect of negation and that the meaning of least is more complex than the meaning of most. A second question concerns whether projection helps clarify the meaning of the in superlatives. Though superlatives appear definite in English (e.g., 'the most books'), in Arabic, the occurrence of 'al' meaning 'the' depends in on word order, which in turn is connected to how far gradability has projected. The research found that when projection passes over a noun, that noun remains indefinite. When projection does not pass over a noun, the noun must become definite. This condition on 'al' does not effect the interpretation of the noun, which in superlative constructions is always semantically indefinite. This means that the definite article does not always signal definiteness, but may serve as a signal of whether projection has taken place. A third question concerns the fact that Arabic 'aktar' can mean both 'most' and 'more', and asks how these two meanings are related. This research found that the meaning of 'aktar' is predictable from the position it occurs in in the sentence, so the differences in interpretation are not entirely internal to 'aktar', but involve its context. The fourth question concerns whether the constraints on projection are the same for aktar when it means 'most' as when it means 'more'. The research found the answer to this question is clearly 'yes', reinforcing the point that the comparative and superlative are two sides of the same coin. This project has substantially advanced our understanding of how gradability is expressed linguistically by capitalizing on unique and unexplored features of an understudied language.