Lauri Karttunen

Regular expressions for language engineering

Many of the processing steps in natural language engineering can be performed using finite state transducers. An optimal way to create such transducers is to compile them from regular expressions. This paper is an introduction to the regular expression calculus, extended with certain operators that have proved very useful in natural language applications ranging from tokenization to light parsing. The examples in the paper illustrate in concrete detail some of these applications.

The Replace Operator

This paper introduces to the calculus of regular expressions a replace operator and defines a set of replacement expressions that concisely encode alternate variations of the operation. Replace expressions denote regular relations, defined in terms of other regular expression operators. The basic case is unconditional obligatory replacement. We develop several versions of conditional replacement that allow the operation to be constrained by context.

Constructing lexical transducers

A lexical transducer, first discussed in Karttunen, Kaplan and Zaenen 1992, is a specialised finite-state automaton that maps inflected surface forms to lexical forms, and vice versa. The lexical form consists of a canonical representation of the word and a sequence of tags that show the morphological characteristics of the form in question and its syntactic category. For example, a lexical transducer for French might relate the surface form veut to the lexical form vouloir+IndPr+SG+P3. In order to map between these two forms, the transducer may contain a path like the one shown in Fig. 1.

Applications of Finite-State Transducers in Natural Language Processing

This paper is a review of some of the major applications of finite-state transducers in Natural Language Processing ranging from morphological analysis to finite-state parsing. The analysis and generation of inflected word forms can be performed efficiently by means of lexical transducers. Such transducers can be compiled using an extended regular expression calculus with restriction and replacement operators. These operators facilitate the description of complex linguistic phenomena involving morphological alternations and syntactic patterns. Because regular languages and relations can be encoded as finite-automata, new languages and relations can be derived from them directly by the finite-state calculus. This is a fundamental advantage over higher-level linguistic formalisms.

Finite-State Non-Concatenative Morphotactics

We describe a new technique for constructing finite-state transducers that involves reapplying the regular-expression compiler to its own output. Implemented in an algorithm called compile-replace, this technique has proved useful for handling non-concatenative phenomena; and we demonstrate it on Malay full-stem reduplication and Arabic stem interdigitation.

Syntax and semantics of questions

This paper presents a novel account of the syntax and semantics of questions, making use of the framework for linguistic description developed by Richard Montague (1974). Certain features of the proposal are based on work by N. Belnap (1963), L. Aqvist (1965), C. L. Baker (1968, 1970), S. Kuno and J. Robinson (1972), C. L. Hamblin (1973), E. Keenan and R. Hull (1973), J. Hintikka (1974), Lewis (1975), and D. Wunderlich (1975), but it differs from all of its predecessors in one way or another. I will start with a number of observations which provide the basis for the treatment of questions presented in the second part of the paper and conclude with a summary and a brief discussion of how the proposed description compares with recent transformational analyses.

Parallel replacement in finite state calculus

This paper extends the calculus of regular expressions with new types of replacement expressions that enhance the expressiveness of the simple replace operator defined in Karttunen (1995). Parallel replacement allows multiple replacements to apply simultaneously to the same input without interfering with each other. We also allow a replacement to be constrained by any number of alternative contexts. With these enhancements, the general replacement expressions are more versatile than two-level rules for the description of complex morphological alternations.

Finite state methods in natural language processing

Finite state methods have been in common use in various areas of natural language processing (NLP) for many years. A series of specialized workshops in this area illustrates this. In 1996, András Kornai organized a very successful workshop entitled Extended Finite State Models of Language. One of the results of that workshop was a special issue of Natural Language Engineering (Volume 2, Number 4). In 1998, Kemal Oflazer organized a workshop called Finite State Methods in Natural Language Processing. A selection of submissions for this workshop were later included in a special issue of Computational Linguistics (Volume 26, Number 1). Inspired by these events, Lauri Karttunen, Kimmo Koskenniemi and Gertjan van Noord took the initiative for a workshop on finite state methods in NLP in Helsinki, as part of the European Summer School in Language, Logic and Information. As a related special event, the 20th anniversary of two-level morphology was celebrated. The appreciation of these events led us to believe that once again it should be possible, with some additional submissions, to compose an interesting special issue of this journal.

Structure Sharing with Binary Trees

Problem: proliferation of copies When words are combined to form phrases, unification is not applied to lexlcat representations directly because it would result in the lexicon being changed. When a word is encountered in a text, a copy is made of its entry, and unification is applied to the copied graph, not the original one. In fact, unification in a typical parser is always preceded by a copying operation. Because of nondeterminism in parsing, it is, in general, necessary to preserve every representation that gets built. The same graph may be needed again when the parser comes back to pursue some yet unexplored option. Our experience suggests that the amount of computational effort that goes into producing these copies is much greater than the cost of unification itself. It accounts for a significant amount of the total parsing time.

Local Textual Inference: Can it be Defined or Circumscribed?

This paper argues that local textual inferences come in three well-defined varieties (entailments, conventional implicatures/presuppositions, and conversational implicatures) and one less clearly defined one, generally available world knowledge. Based on this taxonomy, it discusses some of the examples in the Pascal text suite and shows that these examples do not fall into any of them. It proposes to enlarge the test suite with examples that are more directly related to the inference patterns discussed.