~ ROA Version, 8/2002. Essentially identical to the Tech Report, with new pagination (but the same footnote and example numbering); correction of typos, oversights & outright errors; improved typography; and occasional small-scale clarificatory rewordings. Citation should include reference to this version.

A data-oriented (empiricist) alternative to the currently pervasive (nativist) Principles and Pa-rameters approach to the acquisition of stress assignment is investigated. A similarity-based algorithm, viz. an augmented version of Instance-Based Learning is used to learn the system of main stress assignment in Dutch. In this nontrivial task a comprehensive lexicon of Dutch monomorphemes is used instead of the idealized and highly simplified description of the empirical data used in previous approaches. It is demonstrated that a similarity-based learning method is effective in learning the complex stress system of Dutch. The task is accomplished without the a priori knowledge assumed to pre-exist in the learner in a Principles and Parameters framework. A comparison of the system's behavior with a consensus linguistic analysis (in the framework of Metrical Phonology) shows that ease of learning correlates with decreasing degrees of marked-ness of metrical phenomena. It is also shown that the learning algorithm captures subregularities within the stress system of Dutch that cannot be described without going beyond some of the theoretical assumptions of metrical phonology.

This paper discusses a perceptron model of the learning and assignment of linguistic stress, using data from nineteen human languages. First, we point out some interesting parallels between aspects of the model and the constructs and predictions of metrical phonology, the linguistic theory of stress. Second, we develop a novel analysis of linguistic stress in terms of ease of perceptron-learnability. These two sets of results suggest that simple statistical learning techniques have the potential to complement, and provide computational validation for, abstract theoretical investigations of language. We then examine why such methodologies should be of interest for linguistic theorizing. Our analysis began at a high level by observing inherent characteristics of various stress systems, much as theoretical linguistics does. However, our explanations changed substantially whenwe included a detailed account of the model's processing mechanisms. Our higher-level, theoretical accou...

This article, which emerges from my collaboration with Alan Prince on Prosodic Morphology, will explore the consequences of the Prosodic Morphology Hypothesis for a fairly complete account of the central regularities of canonical form in two Semitic root-and-pattern morphological systems, those of Arabic and Akkadian. We will see that the core of the Arabic nominal system is templatic in character, with templates that conform to (3). But two more specialized nominal constructions depart from (3) and are provably non-templatic. They are analyzed instead in terms of two other notions from Prosodic Morphology theory, prosodic circumscription and a-templatic prosodic morphology. The Arabic and Akkadian verb system is even more radically non-templatic; just a single template underlies all verb forms, and other morphological regularities are derived by rules of affixation, sometimes via prosodic circumscription.

We describe a program for assigning correct stress contours to nominals in English. It makes use of idiosyncratic knowledge about the stress behavior of various nominal types and general knowledge about English stress rules. We have also investigated the related issue of parsing complex nominals in English. The importance of this work and related research to the problem of text-to-speech is 'discussed.

This paper offers a provisional mathemat~ ical typology of metrical representations.

this document was supported by the National Science Foundation (grant SBR9420424)

Explaining how children infer grammatical rules based on their limited finite experience is central to the hypothesis of universal grammar—that collection of inductive principles which allows learning to happen. Because children and languages are complex and many factors influence acquisition—physiological, sociolinguistic, articulatory, perceptual,

vorgelegt von

Prosody, in terms of theories of phonological representation, refers to syllabicity, length, syllable organisation, stress and related concepts, and certain sequencing relations between segment sequences, usually between adjacent syllables. Skeletal positions: The problem. One of the basic questions of prosody is how segmental length should be represented. Length could be treated as a binary feature and is so treated in the SPE theory, analogous to nasality or voicing. However, it has long been recognised that to a considerable extent, long vowels behave the same as two short vowels and long consonants often behave like two short consonants. This fact is so widely recognised that it has become standard for long segments to be written aa, uu, tt, nn and so on. An example of this patterning is found in the fact that in Yawelmani, long vowels are shortened before clusters of two consonants as well as before geminate (long) consonants (see Chapter 7). This rule interacts with an epenthesis rule which inserts i or u (depending on the preceding vowel) between the first two of three consonants. This shortening rule explains why the underlying long vowel of /Üa:ml-al / and /Üa:ml-it / is short on the surface ˜ [Üamlal] ‘help (dubitative)’, [Üaml-it] ‘help (passive aorist)’. Examples like /Üa:ml-hin/, /Üa:ml-k’a / which surface as [Üa:mil-hin] ‘help (nonfuture)’, [Üa:mil-k’a] ‘help