The Dutch anaphoric possessive construction (APC), as exemplified by Tom zijn fiets ‘Tom his bike’, shows a peculiar mix of regularity and idiosyncracy. The article provides a theory-neutral description of its properties and quantitative information about its use in two treebanks, one of spoken Dutch (CGN) and one of written Dutch (Lassy Small). It argues that the APC has a right branching structure and models it in the framework of Constructional Head-driven Phrase Structure Grammar. The latter’s organization of constructions in terms of a finegrained hierarchy of phrase types is shown to provide the means to capture both what the APC has in common with other possessive constructions and what is idiosyncratic of it.

We use Lambek’s pregroups and the framework of compositional distributional models of language (“DisCoCat”) to study translations from Japanese to English as pairs of functors. Adding decorations to pregroups, we show how to handle word order changes between languages.

Many patterns found in natural language syntax have multiple pos-sible explanations or structural descriptions. Even within the cur-rently dominant Minimalist framework (Chomsky 1995, 2000), it is not uncommon to encounter multiple types of analyses for the same phenomenon proposed in the literature. A natural question, then, is whether one could evaluate and compare syntactic proposals from a quantitative point of view. In this paper, we show how an evaluation measure inspired by the minimum description length principle (Rissa-nen 1978) can be used to compare accounts of syntactic phenomena implemented as minimalist grammars (Stabler 1997), and how argu-ments for and against this kind of analysis translate into quantitative differences.

We derive well-understood and well-studied subregular classes of formal languages purely from the computational perspective of algorithmic learning problems. We parameterise the learning problem along dimensions of representation and inference strategy. Of special interest are those classes of languages whose learning algorithms are necessarily not prohibitively expensive in space and time, since learners are often exposed to adverse conditions and sparse data. Learned natural language patterns are expected to be most like the patterns in these classes, an expectation supported by previous typological and linguistic research in phonology. A second result is that the learning algorithms presented here are completely agnostic to choice of linguistic representation. In the case of the subregular classes, the results fall out from traditional model-theoretic treatments of words and strings. The same learning algorithms, however, can be applied to model-theoretic treatments of other linguistic representations such as syntactic trees or autosegmental graphs, which opens a useful direction for future research.

The goal of theory construction is explanation: for language, theory for particular languages (grammar) and for the faculty of language FoL (the innate endowment for language acquisition). A primitive notion of simplicity of grammars is number of symbols, but this is too crude. An improved measure distinguishes grammars that capture genuine properties of language from those that do not. The theory of FoL must meet the empirical conditions of learnability (under extreme poverty of stimulus), and evolvability (given the limited but not insignificant evidence available). Recent work provides promising insights into how these twin conditions may be satisfied.

In this paper, we investigate how the prediction paradigm from machine learning and Natural Language Processing (NLP) can be put to use in computational historical linguistics. We propose word prediction as an intermediate task, where the forms of unseen words in some target language are predicted from the forms of the corresponding words in a source language. Word prediction allows us to develop algorithms for phylogenetic tree reconstruction, sound correspondence identification and cognate detection, in ways close to attested methods for linguistic reconstruction. We will discuss different factors, such as data representation and the choice of machine learning model, that have to be taken into account when applying prediction methods in historical linguistics. We present our own implementations and evaluate them on different tasks in historical linguistics.

We present a model for predicting inflected word forms based on morphological analogies. Previous work includes rule-based algorithms that determine and copy affixes from one word to another, with limited support for varying inflectional patterns. In related tasks such as morphological reinflection, the algorithm is provided with an explicit enumeration of morphological features which may not be available in all cases. In contrast, our model is feature-free: instead of explicitly representing morphological features, the model is given a demo pair that implicitly specifies a morphological relation (such as write:writes specifying infinitive:present). Given this demo relation and a query word (e.g. watch), the model predicts the target word (e.g. watches). To address this task, we devise a character-based recurrent neural network architecture using three separate encoders and one decoder.Our experimental evaluation on five different languages shows tha the exact form can be predicted with high accuracy, consistently beating the baseline methods. Particularly, for English the prediction accuracy is 95.60%. The solution is not limited to copying affixes from the demo relation, but generalizes to words with varying inflectional patterns, and can abstract away from the orthographic level to the level of morphological forms.

A recently proposed balanced-bracket encoding (Yli-Jyrä and GómezRodríguez 2017) has given us a way to embed all noncrossing dependency graphs into the string space and to formulate their exact arcfactored inference problem (Kuhlmann and Johnsson 2015) as the best string problem in a dynamically constructed and weighted unambiguous context-free grammar. The current work improves the encoding and makes it shallower by omitting redundant brackets from it. The streamlined encoding gives rise to a bounded-depth subset approximation that is represented by a small finite-state automaton. When bounded to 7 levels of balanced brackets, the automaton has 762 states and represents a strict superset of more than 99.9999% of the noncrossing trees available in Universal Dependencies 2.4 (Nivre et al. 2019). In addition, it strictly contains all 15-vertex noncrossing digraphs. When bounded to 4 levels and 90 states, the automaton still captures 99.2% of all noncrossing trees in the reference dataset. The approach is flexible and extensible towards unrestricted graphs, and it suggests tight finite-state bounds for dependency parsing, and for the main existing parsing methods.

We introduce algorithms that, given a Finite-State Automaton (FSA), compute a minimal set of forbidden local factors that define a Strictly Local (SL) tight approximation of the stringset recognised by the FSA and the set of forbidden piecewise factors that define a Strictly Piecewise (SP) tight approximation of that stringset, as well as a set of co-SL factors that, together with the SL and SP factors provide a set of purely conjunctive literal constraints defining a minimal superset of the stringset recognised by the automaton.Using these, we have built computational tools that have allowed us to reproduce, by nearly purely computational means, the work of Rogers and his co-workers (Rogers et al. 2012) in which, using a mix of computational and analytical techniques, they completely characterised, with respect to the Local and Piecewise Subregular hierarchies, the constraints on the distribution of stress in human languages that are documented in the StressTyp2 database.Our focus, in this paper, is on the algorithms and the method of their application. The phonology of stress patterns is a particularly good domain of application since, as we show here, they generally fall at the very lowest levels of complexity. We discuss these phonological results here, but do not consider their consequences in depth.

For more than half a century, finite-state methods and mathematical linguistics have benefitted from a close relation and fruitful interaction. Both research fields aim to achieve a deeper understanding of human language by means of mathematical techniques. For the automated processing of language by computers such a mathematical basis is an indispensable prerequisite. Historically, the goal pursued by mathematical linguists to formalize natural language syntax in a computer-accessible way was one of the strongest driving forces behind the development of finite-state methods. Thus, it is no exaggeration to say that the field of finite-state methods owes its existence to a large extent to mathematical linguistics. In turn, continued research on finite-state methods has resulted in a categorization of various kinds of language classes and language aspects, together with efficient and provably correct algorithms, thus expanding our understanding of the mathematical properties of language.