This paper presents a set of mathematical and computational tools for manipulating and reasoning about regular languages and regular relations and argues that they provide a solid basis for computational phonology. It shows in detail how this framework applies to ordered sets of context-sensitive rewriting rules and also to grammars in Koskenniemi's two-level formalism. This analysis provides a common representation of phonological constraints that supports efficient generation and recognition by a single simple interpreter. 1.

this paper are the following: (1) Lexical representations tend to be arbitrary. Because it is difficult to write and test two-level systems that map between pairs of radically dissimilar forms, lexical representations in existing two-level analyzers tend to stay close to the surface forms. This is not a problem for morphologically simple languages like English because, for most words, inflected forms are very similar to the canonical dictionary entry. Except for a small number of irregular verbs and nouns, it is not difficult to create a two-level description for English in which lexical forms coincide with the canonical citation forms found in a dictionary. However, current analyzers for morphologically more complex languages (Finnish and Russian, for example) are not as satisfying in this respect. In these systems, lexical forms typically contain diacritic markers and special symbols; they are not real words in the language. For example, in Finnish the lexical counterpart of otin `I took' might be rendered as

this paper we present a finite-state model of phonology in which automata are the descriptions and tapes (or strings) are the objects being described. This provides the formal semantics for an autosegmental phonology without structure-changing rules. Logical operations on the phonological domain--such as conjunction, disjunction, and negation--make sense since the phonological domain consists of descriptions rather than objects. These operations as applied to automata are the straightforward operations of intersection, union, and complement. If the arrow in a rewrite rule is viewed as logical implication, then a phonological rule can also be represented as an automaton, albeit a less restrictive automaton than would be required for a lexical representation. The model is then compared with the transducer models for autosegmental phonology of Kay (1987), Kornai (1991), and Wiebe (1992). We conclude that the declarative approach to phonology presents an attractive way of extending finite-state techniques to autosegmental phonology while remaining within the confines of regular grammar

This I)aper presents an implemented multi-tape twolevel model capable, of describing Semitic non-linear morphology. The eomputational fi'amcwork behind the eraTent work is motivated by [Kay 1987]; the formal isdn presented here is an extension to the formalism pro'ted by [Puhnan and Hepple 1993]. The objettires of the current work arc: to stay as (',lose as possible, in spirit, to standard two-level morphology, to st,w close to the linguistic description of Semitic stems, and to present a model which can be used with ease by the Semitist. The paper illustrates that if finite-state transducers (FSTs) in a standard two-level mori)hology modal are replaced with multi-tape anxiliary versions (AFSTs), one can account for Semitic root-andq)attern morphology using high level notation.

Arabic inflectional morphology requires infixation, prefixation and suffixation, giving rise to a large space of morphological variation. In this paper we describe an approach to reducing the complexity of Arabic morphology generation using discrimination trees and transformational rules. By decoupling the problem of stem changes from that of prefixes and suffixes, we gain a significant reduction in the number of rules required, as much as a factor of three for certain verb types. We focus on hollow verbs but discuss the wider applicability of the approach.

Phonology may be briefly defined as the study of sound patterns in spoken language. One of the most well-known computational models of phonology, Koskenniemi's two-level phonology, is based on an underlying linguistic theory that has been superseded by autosegmental phonology, which began with the work of Goldsmith. There is a need for computational models that are faithful to this more recent theory. Such a model can form the basis of a computational tool that can quickly and accurately check the validity of a phonological analysis on a large amount of phonetic data, freeing the linguist from the tedious and error-prone task of doing this by hand. This thesis presents a new computational model of phonology that is faithful to standard autosegmental theory, that has clearly adequate expressive power, and that is suitable as the basis for a tool for phonological analysis. It follows on very recent efforts by Kornai and Bird & Ellison to model autosegmental phonology. The model is based ...

This paper describes a method of analysing words through morphological decomposition when the lexicon is incomplete. The method is used within a text-to-speech system to help generate pronunciations of unknown words. The method is achieved within a general morphological analyser system using Koskenniemi two-level rules.

In this report, we present the automatic acquisition of a lexicon involving the use of morphosemantic lexical rules on a core lexicon which was acquired semi-automatically. The process of large-scale dictionary making, as we de ned it to build Spanlex, a Spanish lexicon, for Mikrokosmos, amachine translation (MT) system based on semantics, is at the centre of this report. Our aim in this one-year project has been to acquire a lexicon of about 35,000 word meanings within a computational semantics approach. This implied to automate the task of acquisition as much as possible by developing and implementing tools to help and guide acquisition. Using this approach, we acquired one- fth of Spanlex. The other four- fths were acquired entirely automatically. This paper focuses on the way we acquired this large-scale high quality lexicon by using derivational morpho-semantic rules. In particular, we deal with the discovery, representation, and use of morphosemantic lexical rules (MSLRs) for the automatic acquisition of a large-scale computational lexicon. The analysis is based on a set of MSLRs implemented and tested on the basis of Spanish businessand nance-related corpora. First, we provide the reader with the methodology to build a computational lexicon in the Ecology of Acquisition � second, we present the MSLRs which helped automate the process of acquisition in From Submit to Submitted via Submission: On Lexical

This paper demonstrates how a (multi-tape) two-level formalism can be used to write two-level grammars for Arabic non-linear morphology using a high level, but computationally tractable, notation. Three illustrative grammars are provided based on CV-, moraicand affixational analyses. These are complemented by a proposal for handling the hitherto computationally untreated problem of the broken plural. It will be shown that the best grammars for describing Arabic non-linear morphology are moraic in the case of templatic stems, and affixational in the case of a-templatic stems. The paper will demonstrate how the broken plural can be derived under two-level theory via the `implicit' derivation of the singular. 1 INTRODUCTION Arabic is known amongst computational linguists, in particular computational morphologists, for its highly inflexional morphology. Its root-and-pattern phenomenon has become the prototype for the evaluation of the few non-linear morphological models which have emerged...

This paper presents an implementation of a technique from vi diachronic linguistics, known as the comparative method, for comparing word forms taken from cognate languages in order to reconstruct aspects of the ancestor language from which the languages are derived. The system is applied to data from a group of Tibeto-Burman languages spoken in Nepal. Daelemans, Gillis, and Durieux: The Acquisition of Stress: A Data-Oriented Approach. This paper consists of a rather striking demonstration that an empiricist learning model actually performs better than the nativist 'Principles and Parameters' approach, concerning the task of assigning primary stress to a corpus of around 5,000 Dutch words. Bird and Klein: Phonological Analysis in Typed Feature Systems. This contribution shows how a model of phonology incorporating complex multi-tiered structures can be integrated with a constraint-based grammar of the HPSG variety. Applications to nonconcatenative morphology in Sierra Miwok and deletion in French are given