Martin Plátek

Modeling syntax of free word-order languages: dependency analysis by reduction

This paper explains the principles of dependency analysis by reduction and its correspondence to the notions of dependency and dependency tree. The explanation is illustrated by examples from Czech, a language with a relatively high degree of word-order freedom. The paper sums up the basic features of methods of dependency syntax. The method serves as a basis for the verification (and explanation) of the adequacy of formal and computational models of those methods.

A scale of context sensitive languages: Applications to natural language

The classes of languages Ti are defined, i being a natural number, for which it is proved that Ti is a proper subset of T i+1, and a proper subset of context-sensitive languages, if Ti is the set of all context-free languages and Ti+1 is the set of all possible translations from Ti by means of a so-called linear deterministic pushdown store transducer; the formal mechanism is close enough to the so-called functional generative description, which has already been applied to large subsets of natural languages.

Two-Way Restarting Automata and J-Monotonicity

We introduce restarting automata as two-way automata in order to obtain a more general model which is closer to our linguistic motivations. We study the notion of j-monotonicity to show the advantages of this model. We show that the j-monotonicity can be considered as a degree of non-context-freeness and that it is a robust notion due to the considered models of restarting automata. Some other aspects concerning power and applications of two-way restarting automata are mentioned.

A Taxonomy of Forgetting Automata

Forgetting automata are nondeterministic linear bounded automata whose rewriting capability is restricted as follows: each cell of the tape can only be “erased” (rewritten by a special symbol) or completely “deleted”.

Degrees of non-monotonicity for restarting automata

In the literature various notions of monotonicity for restarting automata have been studied. Here we introduce two new variants of monotonicity for restarting automata and for two-way restarting automata: left-monotonicity and right-left-monotonicity. It is shown that for the various types of deterministic and nondeterministic (two-way) restarting automata without auxiliary symbols, these notions yield infinite hierarchies, and we compare these hierarchies to each other. Further, as a tool used to simplify some of the proofs, the shrinking restarting automaton is introduced, which is a generalization of the standard (length-reducing) restarting automaton to the weight-reducing case. Some of the consequences of this generalization are also discussed.

Forgetting automata and context-free languages

It is shown that context-free languages can be characterized by linear bounded automata with the following restriction: the head can either move right without rewriting or move left with erasing the current cell (i.e. rewriting it with a special, nonrewriteable, symbol). If, instead of erasing, we consider deleting (complete removing of the cell), the corresponding automata are less powerful.

Restarting Automata with Rewriting

Motivated by natural language analysis we introduce restarting automata with rewriting. They are acceptors on the one hand, and (special) regulated rewriting systems on the other hand. The computation of a restarting automaton proceeds in cycles: in each cycle, a bounded substring of the input word is rewritten by a shorter string, and the computation restarts on the arising shorter word.

Monotone deterministic RL-Automata don't need auxiliary symbols

It is known that for monotone deterministic one-way restarting automata, the use of auxiliary symbols does not increase the expressive power. Here we show that the same is true for deterministic two-way restarting automata that are right- or left-monotone. Actually in these cases it suffices to admit delete operations instead of the more general rewrite operations. In addition, we characterize the classes of languages that are accepted by these types of two-way restarting automata by certain combinations of deterministic pushdown automata and deterministic transducers.

A Prototype of a Grammar Checker for Czech

This paper describes the implementation of a prototype of a grammar based grammar checker for Czech and the basic ideas behind this implementation. The demo is implemented as an independent program cooperating with Microsoft Word. The grammar checker uses specialized grammar formalism which generally enables to check errors in languages with a very high degree of word order freedom.

Characterization of Context-Free Languages by Erasing Automata

It is shown that context-free languages are recognizable by (non-deterministic) erasing automata; thereby a hypothesis of [1] is denied. In addition, the class of context-free languages is characterized by means of the automata which erase each cell at the second visit at latest.