This dissertation is intended as a contribution to the study and understanding of how language is used among the Akawaio peoples of Guyana, South America. It is a first attempt, limited to one Amerindian village, Waramadong, a community that has not been studied very much. The study begins by analysing four Akawaio speech genres, tareng 'ritual healing chant', mire aburobodi 'praising rhymes for children', pandong 'story', and zegareme'no 'personal narratives', at the levels of both content and grammar. The first entails making anthropological, ethnographical, sociolinguistic and general cultural commentaries on the content and social context within which speech is performed. This is to investigate how the Akawaio speech genres are categorised and classified and in what ways they are performed, interrelated and connected to the wider domain of speaking in Akawaio society. The second is committed to analysing the observed linguistic variation in the speech genres at the levels of both dialectal variation across speakers and stylistic variation across the genres. The analysis takes a multidisciplinary approach, transcending traditional linguistic boundaries, and invoking both social and linguistic theories, especially in the analysis of Akawaio spirituality as a crucial component to understanding the native Akawaio view of speech genres. Thus, this study offers a primary description of a wider, extended view of what is known about language and culture in Akawaio society. This dissertation also seeks to rectify a serious situation, to provide an emphatic counterexample to the common image in linguistics and anthropological literature, where Amerindian communities are treated as largely homogenous groups. The premise behind most anthropological and linguistic studies is that everyone acts and speaks alike within these societies. One aim of this study is to replace this homogenized image of the Amerindian with a richer, more complex and internally diverse picture, of the kind shown here for Waramadong. Appendix B presents a small, but representatively diverse, selection of transcribed, translated, and linguistically annotated texts, representing a small subset of the overall collection of texts recorded for this study.

In Creek, a Muskogean language, nominalization is formally signalled by a loss of inflectional morphology and the occurrence of derivational morphology. A nominalization may be taken to be a concrete interpretation of the event itself or one of its participants. The verbal derivational morpheme $\{$ip$\}$ signals medio-passive voice, in which the executor of the event is not the agent of the event. The verbal derivational morpheme $\{$ec$\}$ signals increased transitivity, by which transitive verbs are derived from intransitive, transitive verbs are made more transitive through an increase of some parameter of transitivity, and causatives are created with the help of the medio-passive morpheme. The middle-voice $\{$k$\}$ morpheme signals that the executor of the event is affected by the action of the event, as in statives, intransitives, and reflexives. Participant agreement type is lexically marked for verbs, but paradigmatic contrast shows the markers to be semantically motivated. Types I and II marking vary with respect to control of the event, and Types II and III marking vary with respect to envelopment by the event. When the dependent verb of a modificational clause is non-tensed, the $\{$ii$\}$ and $\{$aa$\}$ suffixes differentiate non-identifiable from identifiable participants, respectively. When the dependent verb is tensed, the $\{$ii$\}$ and $\{$aa$\}$ suffixes differentiate mentioned events from asserted events, respectively. The semantic connection between the two uses of $\{$ii$\}$ and $\{$aa$\}$ are backgrounding and foregrounding, respectively. Non-identifiable participants and mentioned events are united in backgrounding and are suffixed with $\{$ii$\}$. Identifiable participants and asserted events are united in foregrounding and are suffixed with $\{$aa$\}$. $\{$t$\}$ and $\{$n$\}$ signal foregrounding and backgrounding, respectively, within the proposition; that is, they determine how a participant or event is foregrounded or backgrounded with respect to other participants or events within the same proposition. The $\{$ooM$\}$ suffix backgrounds participants and events with respect to other propositions, as in answering questions, or with respect to the ontology of the participant or event itself.

The Warihio language is a member of the Uto-Aztecan family. The language consists of two dialects: the Upland Warihio in the mountains of Chihuahua and the River Warihio along the Mayo River in Sonora, Mexico. With the various Tarahumara dialects, and Yaqui and Mayo languages, it makes up the Taracahitic sub-group of the Sonoran branch of the Uto-Aztecan family of languages. All of the field and supporting data for this work comes from the River dialect. This work deals with all of the major linguistic aspects of the River Warihio language, including a brief description of its phonology, major and minor word classes, noun phrase, relative clauses, simple sentence structure, negation, voice, and complex sentences structure. Likewise, a short comparative section within Uto-Aztecan languages of some relevant aspects of the Warihio grammar. Also included is a basic Warihio-English-Spanish dictionary and several analyzed texts. These appendixes provide natural language data for study of areas not covered in detail here. Chapter one provides information regarding ethnographical aspects of the Warihio people; it also establishes the phonemic inventory of the language and the notational system used through the dissertation. In chapter one I also propose a stress pattern based in the information about possible combination of roots and affixes allowed in the language. The main theorethical-typological contributions that the study of Warihio might provide are contained in the following chapters: Chapter 5: Simple sentence. Flexibility in order constituent displayed by Warihio texts and its relation with the focus phenomena are described in this chapter. Coding and control properties as well as participants behaviour are also described here. Chapter 7: Voice. I have integrated different voice phenomena such as passive, causative, reflexive, applicative, external possession, and ethical dative in a general semantic frame of voice. I describe typologically interesting findings in the passive and causative constructions. River Warihio has some interesting contrasting aspects within Uto-Aztecan family and morpho-syntactic features that are relevant theoretico-typologically. Its flexible pragmatically motivated constituent order altogether with the lack of coding properties for grammatical relations make Warihio an unusual language within Uto-Aztecan family and cross-linguistically as well.

The semantics of determiners in field data from two Philippine languages, Ilokano and Yogad, is characterized and compared. In Ilokano, this content appears as gradations along a cline of "individuation." In Yogad, the semantics represents successive degrees of "actualization." In both languages, the function of this semantics is to form and delineate participants by segregating these from the ground of quality and event and also to orient within an existing matrix of knowledge the participants thus formed. The name "determinacy" is given to this participant-forming semantics as a means of comparing it across languages. Determinacy, as exemplified in Ilokano (individuation) and Yogad (actualization), is motivated by the cognitive principle FOCUSSED--DIFFUSE. This principle is inherent in the process by which variance in focal attention organizes the continuum of the cognitive experience of an organism. Variable focal attention is the cognitive-psychological basis for determinacy and, therefore, for participant formation in language. The operation of the FOCUSSED--DIFFUSE principle in connection with focal attention outside of language is illustrated in human vision and visual perception and in sonar echolocation in bats. Because the FOCUSSED--DIFFUSE principle is a cognitive universal and is a parameter of meaning characteristic of intelligence itself, we conclude that determinacy is also a linguistic universal, i.e., that it is a constant presence in language, even in languages which lack determining forms. In proposing a cognitive motivation for determinacy, this study challenges the privileging of discourse pragmatics in recent attempts to understand the function of determiners. The analysis of the Ilokano and Yogad data shows that in these languages determiners are not involved in the management of information flow in connected discourse. The study rejects the notion of the modularity of language or of linguistic intelligence; it argues that determinacy in language cannot be adequately described apart from understanding the way in which the FOCUSSED--DIFFUSE principle operates in other cognitive domains.

Comparative and classificatory studies of Cariban languages, despite their long history (starting with Gilij in 1782), have been few and unsatisfactory, mainly due to the lack of necessary documentation of the languages in question. Based on a large amount of new descriptive data, as well as on published sources, the present work attempts to demonstrate the closer genetic relationship between a subgroup of three Cariban languages, Akuriyo, Tiriyo, and Karihona, the last two of which were considered to belong to very distant branches of the family in a still widely cited classification (Durbin 1977). This demonstration takes the form of a reconstruction of the main aspects of the segmental phonology and inflectional morphology (person, number, evidentiality, tense/aspect/mood) of the proto-language, which I propose to call Proto-Taranoan. A preliminary etymological dictionary, as well as some remarks on the history of the speakers, is also included.

In this dissertation I examine first how the dynamicity of prepositions and the meaning of verbs interact in English motion constructions, and second what role the constructions themselves play in this interaction. I adopt a Construction Grammar perspective, which assumes that constructions can contribute to the interpretation of utterances. Based on an extensive corpus study and additional evidence from a survey on acceptability judgments, I investigate the limits of the semantic import of constructions with respect to the expression of PATH. To that end, I determine the degree of dynamicity of 19 prepositions on the basis of their frequency of occurrence in static versus dynamic utterance types. This degree of dynamicity measures a preposition's contribution to the semantics of utterances instantiating the Caused-Motion Construction (CMC) or the Intransitive-Motion Construction (IMC) with respect to the predication of a dynamic PATH. I show that a dynamic PATH needs to be lexically expressed in motion constructions, either by the preposition or by the verb, if it is not retrievable through contextual factors. This has direct repercussions for the use of prepositions in the two constructions. I show that coercion of non-dynamic prepositions into a dynamic interpretation is only possible if both the verb and the preposition inherently profile the endpoint. The power of constructions to coerce the meanings of lexical items, i.e. to influence the canonical interpretation of these items where they do not correspond to the construction's semantics, is thus much more restricted than commonly assumed. I also show that constructions are limited with respect to the number of lexical items that they can coerce in a given instantiation. The CMC and the IMC can either coerce a non-motion verb or a non-dynamic preposition, but the combination of non-motion verbs with non-dynamic prepositions yields utterances that no longer express a directional PATH. Overall, this thesis illustrates that it is indispensable to consider the lexical semantics of the instantiating verbs and their co-occurrence restrictions with specific prepositions, even when examining the workings of more abstract linguistic units such as constructions.

This dissertation is a synchronic description of the grammar of the Matses language (also know as Mayoruna; Panoan family) as currently spoken by the Matses people living in Amazonian Peru and Brazil. The Matses language is spoken by 2000--2200 people, Amerindians who were first contacted in 1969 and continue to pursue traditional subsistence practices. This is the first attempt at a comprehensive description of the grammar of Matses; full-length grammars of no other Panoan language exist. Matses phonology, morphology, and syntax are the principal topics of this work. It follows a traditional format and is organized so that it can be used as a reference. The introductory chapter briefly provides information about classification of the language (particularly the Mayoruna subgroup), demography, physical setting, history, ethnography, literature review, and methodology. The second chapter describes Matses phonology, including an inventory of distinctive segments, syllable structure, morphophonology, prosody, sound symbolism, and borrowing. The next seven chapters are on morphology (an introduction to morphology, followed by six chapters describing the morphology of nouns, verbs, adjectives, adverbs, postpositions, and particles). All aspects of morphology are treated in these sections, including identification of word classes and subclasses, affixation, clitics, reduplication, and class-changing processes. The last three chapters are on syntax (phrases, one-clause sentences, and multi-clause sentences). The appendix contains three parsed texts. Matses has six vowels and 15 consonants. A word-level alternating rhythmic stress pattern characterizes the sound of the language. Morphologically, Matses stands between isolating and polysynthetic languages, and between agglutinative inflecting/fusional languages. It is the large number of morphological possibilities that is striking about Matses, not the length of its words. Interesting morphological properties include a complex system for coding evidentiality, an elaborate system of directional verbal suffixes, and adverb transitivity agreement. Constituent order is essentially free from syntactic restrictions. Subordination is achieved through expansion of syntactic slots though class-changing processes. Clause-chaining is a prominent feature of Matses discourse with sentences of up to ten clauses. Interesting syntax includes ergative-absolutive case marking alongside nominative-accusative person agreement, and three-place verbs with identical objects.

This thesis is an examination of the regional dialectal influence of colloquial varieties of Arabic on the allophony of vowels in Modern Standard Arabic. It is an expansion of research done by Al-Ani (1970), who claims that speakers' native, colloquial dialect of Arabic has an influence upon the realization of phonological patterns in standardized MSA, and that these influences can be seen to differ in small but noticeable and structured ways. This thesis lends support to Al-Ani's research on Iraqi Arabic by examining the dialects spoken in Lebanon and the United Arab Emirates. Analysis of the data obtained was conducted under the methodology of Articulatory Phonology (Browman & Goldstein, 1992), a theory of phonology which examines phonological patterns in terms of articulatory gestures. This thesis shows that Articulatory Phonology is a successful method of analysis, that not only describes phonological patterning, but explains it in light of articulatory gestures.

The topic of naming and classification reveals human language as a construct of infinite variability. Not only do naming and classification occur in everyday life, but especially within language investigation they document an indispensable axiomatic basic and serve as a tool for linguists and as the main focus of their study. De Saussure, as the founder of structuralism, focuses his language analysis on structural elements and thus lays the foundation for modern linguistic theory. The "cone" is posited as a figurative representation to explain the connection between the different fields of linguistics and to reveal their unifying central principle. Not only can this construction be used to reflect traditional linguistic terminology, such as diachronic and synchronic, but it also provides a means for portraying a horizontal-vertical language analysis within inter-lingual, as well as intra-lingual investigation in various fields of linguistics. In conclusion, linguistics is shown to be a science based throughout upon classificatory models and their names.

En el siglo XVI la letra procesal degenero en Espana y America volviendose ilegible debido al exceso de abreviaturas, nivelacion de las grafias, descuido al trazar sus rasgos y falta de separacion de las palabras; en sintesis, se transformo en la procesal encadenada. Al llegar a este estado el proceso recede y la caligrafia espanola ingresa, tardiamente y con caracteristicas peculiares, a las corrientes renovadoras que imperaban en el resto de Europa. El resultado de esta evolucion es lo que conocemos como letra bastarda. La escritura refleja las dudas linguisticas de los hablantes enfrentados a la falta de correspondencia entre sonidos y grafias, la necesidad de expresar nuevos conceptos o renovar los existentes y la influencia de sistemas sintacticos diversos. Los seis documentos seleccionados para esta tesis ejemplifican significativos rasgos paleograficos y linguisticos que caracterizaron estos procesos evolutivos en el siglo XVII.