Syntactic Class Research Articles

Tradeoffs in the complexity of backdoors to satisfiability: dynamic sub-solvers and learning during search

There has been considerable interest in the identification of structural properties of combinatorial problems that lead to efficient algorithms for solving them. Some forms of structure, such as properties of the underlying constraint graph, are "easily" identifiable. Others, such as backdoor sets, are of interest because they capture key aspects of state-of-the-art constraint solvers as well as of many real-world problem instances. While backdoors were originally introduced to capture propagation based simplification mechanisms of solvers, they have recently been studied also in the context of tractable syntactic classes such as 2CNF and Horn. These syntactic classes, however, do not capture key aspects of solvers such as empty clause (i.e., violated constraint) detection. We show that incorporating inconsequential sounding features such as empty clause detection has a dramatic impact on both the complexity of finding a backdoor of size k (which becomes harder in the worst case) and on the size of the resulting backdoor (which can become arbitrarily smaller). Empirically, we show that commonly employed polynomial-time "dynamic" constraint propagation mechanisms, such as unit propagation, pure literal elimination, and probing, often lead to much smaller backdoor sets in real-world domains than "statically" defined classes such as Horn and RHorn, thereby capturing structure much more succinctly. We also reveal the inherent limits of the simpler concept of deletion backdoors, specifically by looking at renamable Horn sub-formulas. Finally, we extend the notion of backdoors to incorporate learning during search--a key aspect of nearly all state-of-the-art systematic SAT solvers--and show, both theoretically and empirically, that this drastically reduces the size of the resulting backdoor set. Our results suggest that structural notions explored for designing efficient algorithms for combinatorial problems should capture both statically and dynamically identifiable properties of the combinatorial problem being solved.

Phonological iconicity.

The arbitrariness of the linguistic sign is a fundamental assumption in modern linguistic theory. In recent years, however, a growing amount of research has investigated the nature of non-arbitrary relations between linguistic sounds and semantics. This review aims at illustrating the amount of findings obtained so far and to organize and evaluate different lines of research dedicated to the issue of phonological iconicity. In particular, we summarize findings on the processing of onomatopoetic expressions, ideophones, and phonaesthemes, relations between syntactic classes and phonology, as well as sound-shape and sound-affect correspondences at the level of phonemic contrasts. Many of these findings have been obtained across a range of different languages suggesting an internal relation between sublexical units and attributes as a potentially universal pattern.

Model Counting in the 2mu-3MON Syntactic Class

The counting model problem in Boolean formulas is #P-complete, i.e., there is no known deterministic algorithm in the classical computability model (Turing machine) that makes this count in polynomial time. The difficulty persists even imposing more restrictive conditions on the syntactic clases of Boolean formulas. In this paper we present a treatable family within the syntactical class 2_-3MON. The identification of this family is done by using the hypergraph associated with simple structures such as chains and cycles. Then, matrix operators acting over these structures are identified; these operators lead to efficient algorithms that perform the model counting on the identified family in linear time for the number of clauses in the instantiated formula; unlike hypergraphic invariant based methods (such as tree width), which perform the count in cubic time.

From Modal to Multimodal Ambiguities: a Classification Approach

This paper deals with classifying ambiguities for Multimodal Languages. It evolves the classifications and the methods of the literature on ambiguities for Natural Language and Visual Language, empirically defining an original classification of ambiguities for multimodal interaction using a linguistic perspective. This classification distinguishes between Semantic and Syntactic multimodal ambiguities and their subclasses, which are intercepted using a rule-based method implemented in a software module. The experimental results have achieved an accuracy of the obtained classification compared to the expected one, which are defined by the human judgment, of 94.6% for the semantic ambiguities classes, and 92.1% for the syntactic ambiguities classes.

On the stable model semantics for intensional functions

AbstractSeveral extensions of the stable model semantics are available to describe ‘intensional’ functions—functions that can be described in terms of other functions and predicates by logic programs. Such functions are useful for expressing inertia and default behaviors of systems, and can be exploited for alleviating the grounding bottleneck involving functional fluents. However, the extensions were defined in different ways under different intuitions. In this paper we provide several reformulations of the extensions, and note that they are in fact closely related to each other and coincide on large syntactic classes of logic programs.

Syntaktische Nebensatzklassen im Deutschen

This article gives an overview of the syntactic classes of subordinate clauses in German. Their functions and structures are described with the help of the model of linear positions and fields. Three questions are treated in more detail: the functions of (i) pronouns and (ii) comparative particles in relation to subordinate clauses and (iii) the role of subordinate clauses in cleft constructions. The article can be used in academic teaching and as a starting point for studies in comparative syntax.

Relational transducers for declarative networking

Motivated by a recent conjecture concerning the expressiveness of declarative networking, we propose a formal computation model for “eventually consistent” distributed querying, based on relational transducers. A tight link has been conjectured between coordination-freeness of computations, and monotonicity of the queries expressed by such computations. Indeed, we propose a formal definition of coordination-freeness and confirm that the class of monotone queries is captured by coordination-free transducer networks. Coordination-freeness is a semantic property, but the syntactic class of “oblivious” transducers we define also captures the same class of monotone queries. Transducer networks that are not coordination-free are much more powerful.

A Statistical Investigation into the Cross-Linguistic Distribution of Mass and Count Nouns: Morphosyntactic and Semantic Perspectives

We collected a database of how 1,434 nouns are used with respect to the mass/count distinction in six languages; additional informants characterized the semantics of the underlying concepts. Results indicate only weak correlations between semantics and syntactic usage. In five out of the six languages, roughly half the nouns in the database are used as pure count nouns in all respects; the other half differ from pure counts over distinct syntactic properties, with fewer nouns differing on more properties, and typically very few at the pure mass end of the spectrum. Such a graded distribution is similar across languages, but syntactic classes do not map onto each other, nor do they reflect, beyond weak correlations, semantic attributes of the concepts. Considerable variability is seen even among speakers of the same language. These findings are in line with the hypo-thesis that much of the mass/count syntax emerges from language- and even speaker-specific grammaticalization.

The Role of Syntactic Class, Frequency, and Word Order in Looking up English Multi-Word Expressions

Multi-word lexical units, such as compounds and idioms, are often problematic for lexicographers. Dictionaries are traditionally organized around single orthographic words, and so the question arises of where to place such complex lexical units. The user-friendly answer would be to include them primarily under the word which users are most likely to look up. But how do we know which words are likely to be looked up? The present study addresses this question by examining the roles of part of speech, word frequency, and word position in guiding the decisions of Polish learners of English as to which component word of a multi-word expression to look up in the dictionary. The degree of word frequency is found to be the strongest predictor, with less frequent words having a significantly greater chance of being selected for consultation. Then there is an independent part of speech-related preference for nouns, with adjectives being second, followed by verbs in third place. Words belonging to the remaining syntactic categories (adverbs, prepositions, conjunctions, determiners, and pronouns) are hardly looked up at all. However, word placement within the multi-word expression does not seem to matter much. This study has implications for dictionary makers in considering how to list multi-word-expressions.

Recycling and replacement repairs as self-initiated same-turn self-repair strategies in Hungarian

The aim of this paper is to explore recycling and replacement repairs as self-initiated same-turn self-repair strategies in Hungarian. The study concentrates on four factors: repair operation types, syntactic class and length of the repaired segment, and site of initiation. In accordance with previous works focusing on the relationship between grammar and repair, the analysis shows that the main organizer of the self-repair process is the speaker's interactional aim. This interactional aim is realized in the interactional functions of repair operations: providing the speaker with extra time in the case of recycling repairs, or exchanging an unintended item in the case of replacement repairs. The working of these interactional functions, however, always adapts to the grammatical possibilities of the particular language. I have assumed that the ratio of early and late initiations in recycling repairs depends on the typical orders of function and content words in languages, i.e. the exploitability of the delaying function of function word recycling. While the recycling replacement ratio is likely to be approximately universally constant, within recycling repairs the ratio of early and late initiations depends on the morphosyntactic structures of languages. As in Hungarian function words usually precede their respective content words, recyclings tend to be initiated after recognizable completion.

Object-oriented programming with gradual abstraction

We describe an experimental object-oriented programming language, ASL2, that supports program development by means of a series of abstraction steps. The language allows immediate object construction, and it is possible to use the constructed objects for concrete problem solving tasks. Classes and class hierarchies can be derived from the objects - via gradual abstraction steps. We introduce two levels of object classification, called weak and strong object classification. Strong object classification relies on conventional classes, whereas weak object classification is looser, and less restrictive. As a central mechanism, weakly classified objects are allowed to borrow methods from each other. ASL2 supports class generalization, as a counterpart to class specialization and inheritance in mainstream object-oriented programming languages. The final abstraction step discussed in this paper is a syntactical abstraction step that derives a source file with a syntactical class form.

Towards a canonical classical natural deduction system

This paper studies a new classical natural deduction system, presented as a typed calculus named λ̲μlet. It is designed to be isomorphic to Curien and Herbelinʼs λ¯μμ˜-calculus, both at the level of proofs and reduction, and the isomorphism is based on the correct correspondence between cut (resp. left-introduction) in sequent calculus, and substitution (resp. elimination) in natural deduction. It is a combination of Parigotʼs λμ-calculus with the idea of “coercion calculus” due to Cervesato and Pfenning, accommodating let-expressions in a surprising way: they expand Parigotʼs syntactic class of named terms.This calculus and the mentioned isomorphism Θ offer three missing components of the proof theory of classical logic: a canonical natural deduction system; a robust process of “read-back” of calculi in the sequent calculus format into natural deduction syntax; a formalization of the usual semantics of the λ¯μμ˜-calculus, that explains co-terms and cuts as, respectively, contexts and hole-filling instructions. λ̲μlet is not yet another classical calculus, but rather a canonical reflection in natural deduction of the impeccable treatment of classical logic by sequent calculus; and Θ provides the “read-back” map and the formalized semantics, based on the precise notions of context and “hole-expression” provided by λ̲μlet.We use “read-back” to achieve a precise connection with Parigotʼs λμ, and to derive λ-calculi for call-by-value combining control and let-expressions in a logically founded way. Finally, the semantics Θ, when fully developed, can be inverted at each syntactic category. This development gives us license to see sequent calculus as the semantics of natural deduction; and uncovers a new syntactic concept in λ¯μμ˜ (“co-context”), with which one can give a new definition of η-reduction.

A Canonical Locally Named Representation of Binding

This paper is about completely formal representation of languages with binding. We have previously written about a representation following an approach going back to Frege, based on first-order syntax using distinct syntactic classes for locally bound variables vs. global or free variables (Sato and Pollack, J Symb Comput 45:598---616, 2010). The present paper differs from our previous work by being more abstract. Whereas we previously gave a particular concrete function for canonically choosing the names of binders, here we characterize abstractly the properties required of such a choice function to guarantee canonical representation, and focus on the metatheory of the representation, proving that it is in substitution preserving isomorphism with the nominal Isabelle representation of pure lambda terms. This metatheory is formalized in Isabelle/HOL. The final section outlines a formalization in Matita of a challenging language with multiple binding and simultaneous substitution. The Isabelle and Matita proof files are available online.

Acquisition of Turkish grammatical morphology by children with developmental disorders

Many children with specific language impairment, Down syndrome or autism spectrum disorder have difficulty learning grammatical morphology, especially forms associated with the verb phrase. However, except for Hebrew, the evidence thus far has come from Indo-European languages. This study investigates the acquisition of grammatical morphology by Turkish-speaking children with developmental disorders. Syntactic, perceptual and usage features of this non-Indo-European language were predicted to lead to patterns of atypical learning that would challenge and broaden current views. Language samples were collected from 30 preschoolers learning Turkish: ten with developmental disorders, ten matched by age and ten by length of utterance. T-SALT then generated mean length of utterance, the total number of noun errors, the total number of verb errors and the per cent use in obligatory contexts for noun suffixes. Analyses also looked at the potential effects of input frequency on order of acquisition. Turkish children in the MLU-W control group, aged 3;4, used noun and verb suffixes with virtually no errors. Children in the group with atypical language showed more, and more persistent, morphological errors than either age or language peers, especially on noun suffixes. Children in the ALD and MLU-W groups were acquiring noun case suffixes in an order that is strongly related to input frequencies. These findings seem to reflect the influence of salience, regularity and frequency on language learning. Typical child-adult discourse patterns as well as the canonical SOV Turkish word order make verb suffixes perceptually salient, available in working memory and frequently repeated. The findings support the view that the language patterns seen in children with atypical development will differ from one language type to the next. They also suggest that regardless of language or syntactic class, children will have greater difficulty with those features of grammar that have higher cognitive processing costs.

The representational adequacy of Hybrid

The Hybrid system (Ambler et al. 2002b), implemented within Isabelle/HOL, allows object logics to be represented using higher order abstract syntax (HOAS), and reasoned about using tactical theorem proving in general, and principles of (co)induction in particular. The form of HOAS provided by Hybrid is essentially a lambda calculus with constants.Of fundamental interest is the form of the lambda abstractions provided by Hybrid. The user has the convenience of writing lambda abstractions using names for the binding variables. However, each abstraction is actually a definition of a de Bruijn expression, and Hybrid can unwind the user's abstractions (written with names) to machine friendly de Bruijn expressions (without names). In this sense the formal system contains a hybrid of named and nameless bound variable notation.In this paper, we present a formal theory in a logical framework, which can be viewed as a model of core Hybrid, and state and prove that the model is representationally adequate for HOAS. In particular, it is the canonical translation function from λ-expressions to Hybrid that witnesses adequacy. We also prove two results that characterise how Hybrid represents certain classes of λ-expression.We provide the first detailed proof to be published that proper locally nameless de Bruijn expressions and α-equivalence classes of λ-expressions are in bijective correspondence. This result is presented as a form of de Bruijn representational adequacy, and is a key component of the proof of Hybrid adequacy.The Hybrid system contains a number of different syntactic classes of expression, and associated abstraction mechanisms. Hence, this paper also aims to provide a self-contained theoretical introduction to both the syntax and key ideas of the system. Although this paper will be of considerable interest to those who wish to work with Hybrid in Isabelle/HOL, a background in automated theorem proving is not essential.

Syntactic Categorization in French-Learning Infants.

Recent work showed that infants recognize and store function words starting from the age of 6-8 months. Using a visual fixation procedure, the present study tested whether French-learning 14-month-olds have the knowledge of syntactic categories of determiners and pronouns, respectively, and whether they can use these function words for categorizing novel words to nouns and verbs. The prosodic characteristics of novel words stimuli for noun versus verb uses were balanced. The only distinguishing cue was the preceding determiners versus subject pronouns, the former being the most common for nouns and the latter the most common for verbs, i.e., Det + Noun, Pron + Verb. We expected that noun categorization may be easier than verb categorization because the co-occurrence of determiners with nouns is more consistent than that of subject pronouns with verbs in French. The results showed that infants grouped the individual determiners as one common class, and that they appeared to use the determiners to categorize novel words into nouns. However, we found no evidence of verb categorization. Unlike determiners, pronouns were not perceived as a common syntactic class.

Automatic Attribution of Quoted Speech in Literary Narrative

We describe a method for identifying the speakers of quoted speech in natural-language textual stories. We have assembled a corpus of more than 3,000 quotations, whose speakers (if any) are manually identified, from a collection of 19th and 20th century literature by six authors. Using rule-based and statistical learning, our method identifies candidate characters, determines their genders, and attributes each quote to the most likely speaker. We divide the quotes into syntactic classes in order to leverage common discourse patterns, which enable rapid attribution for many quotes. We apply learning algorithms to the remainder and achieve an overall accuracy of 83%.

Speech errors of amnesic H.M.: Unlike everyday slips-of-the-tongue

Three language production studies indicate that amnesic H.M. produces speech errors unlike everyday slips-of-the-tongue. Study 1 was a naturalistic task: H.M. and six controls closely matched for age, education, background and IQ described what makes captioned cartoons funny. Nine judges rated the descriptions blind to speaker identity and gave reliably more negative ratings for coherence, vagueness, comprehensibility, grammaticality, and adequacy of humor-description for H.M. than the controls. Study 2 examined “major errors”, a novel type of speech error that is uncorrected and reduces the coherence, grammaticality, accuracy and/or comprehensibility of an utterance. The results indicated that H.M. produced seven types of major errors reliably more often than controls: substitutions, omissions, additions, transpositions, reading errors, free associations, and accuracy errors. These results contradict recent claims that H.M. retains unconscious or implicit language abilities and produces spoken discourse that is “sophisticated,” “intact” and “without major errors.” Study 3 examined whether three classical types of errors (omissions, additions, and substitutions of words and phrases) differed for H.M. versus controls in basic nature and relative frequency by error type. The results indicated that omissions, and especially multi-word omissions, were relatively more common for H.M. than the controls; and substitutions violated the syntactic class regularity (whereby, e.g., nouns substitute with nouns but not verbs) relatively more often for H.M. than the controls. These results suggest that H.M.’s medial temporal lobe damage impaired his ability to rapidly form new connections between units in the cortex, a process necessary to form complete and coherent internal representations for novel sentence-level plans. In short, different brain mechanisms underlie H.M.’s major errors (which reflect incomplete and incoherent sentence-level plans) versus everyday slips-of-the tongue (which reflect errors in activating pre-planned units in fully intact sentence-level plans). Implications of the results of Studies 1–3 are discussed for systems theory, binding theory and relational memory theories.

Syntactic classes of the Arabic passive participle: And how they should be rendered into English

The main concern of this article is to provide an analysis of the syntactic classes of Arabic passive participle forms and discuss their translations based on a comparative study of two English Quranic translations by Ali (1934) and Pickthall (1930). The study attempts to answer two questions: (a) Should we translate the Arabic passive participle into an English nominal, verbal, adjectival or adverbial? and (b) What are the factors that determine the choice of one translation or the other? So, it compares the two translations to analyze the different English translations of the Arabic passive participle. A corpus of 350 sentences has been randomly selected from the source text, together with their 700 translations in the target texts. The two translations of all the sentences are compared and analyzed in terms of syntactic and semantic features. The various English translations of the Arabic passive participle forms are presented with a count of the examples representing them in the corpus and their percentages. Then, the contextual reference of each translation is studied and accounted for.

First Person Singular

We use digital library technology to help language learners express themselves by capitalizing on the humangenerated text available on the Web. From a massive collection of n-grams and their occurrence frequencies we extract sequences that begin with the word I, sequences that begin a question, and sequences containing statistically significant collocations. These are preprocessed, filtered, and organized as a digital library collection using the Greenstone software. Users can search the collection to see how particular words are typically used and browse by syntactic class. The digital library is richly interconnected to other resources. It includes links to external vocabularies and thesauri so that users can retrieve words related to any term of interest, and links the collection to the web by locating sample sentences containing these patterns and presenting them to the user. We have conducted an evaluation of how useful the system is in helping students, and the impact it has on their writing. Finally, language activities generated from the digital library content have been designed to help learners master important emotion related vocabulary and expressions. We predict that the application of digital library technology to assist language students will revolutionize second language learning.