## Contrasting applications of logic in natural language syntactic description (2005)

Venue: | Logic, Methodology and Philosophy of Science: Proceedings of the Twelfth International Congress |

Citations: | 10 - 1 self |

### BibTeX

@INPROCEEDINGS{Pullum05contrastingapplications,

author = {Geoffrey K. Pullum and Barbara C. Scholz},

title = {Contrasting applications of logic in natural language syntactic description},

booktitle = {Logic, Methodology and Philosophy of Science: Proceedings of the Twelfth International Congress},

year = {2005},

pages = {481--503},

publisher = {King’s College Publications}

}

### OpenURL

### Abstract

Abstract. Formal syntax has hitherto worked mostly with theoretical frameworks that take grammars to be generative, in Emil Post’s sense: they provide recursive enumerations of sets. This work has its origins in Post’s formalization of proof theory. There is an alternative, with roots in the semantic side of logic: model-theoretic syntax (MTS). MTS takes grammars to be sets of statements of which (algebraically idealized) well-formed expressions are models. We clarify the difference between the two kinds of framework and review their separate histories, and then argue that the generative perspective has misled linguists concerning the properties of natural languages. We select two elementary facts about natural language phenomena for discussion: the gradient character of the property of being ungrammatical and the open nature of natural language lexicons. We claim that the MTS perspective on syntactic structure does much better on representing the facts in these two domains. We also examine the arguments linguists give for the infinitude of the class of all expressions in a natural language. These arguments turn out on examination to be either unsound or lacking in empirical content. We claim that infinitude is an unsupportable claim that is also unimportant. What is actually needed is a way of representing the structure of expressions in a natural language without assigning any importance to the notion of a unique set with definite cardinality that contains all and only the expressions in the language. MTS provides that.

### Citations

507 |
The language instinct
- Pinker
- 1994
(Show Context)
Citation Context ...s that of (10a), then (10b) does not violate any constraint at all. Introductory works on language and linguistics commonly observe that expressions containing pseudo-words are well formed. Many (see =-=Pinker 1994-=-:89 for a typical example) reprint the first stanza of Lewis Carroll’s Jabberwocky (“’Twas brillig, and the slithy toves / Did gyre and gimble in the wabe; / All mimsy were the borogoves...”) to make ... |

476 |
Automata on infinite objects
- Thomas
- 1990
(Show Context)
Citation Context ...essions have without saying anything about how big expressions can be. 14 The relevant definitions can easily be modified, and have been for theoretical reasons within logic and computer science (see =-=Thomas 1990-=-); but the standard definitions absolutely exclude infinite expressions. 15 The observation that model-theoretic (“nonconstructive”) grammars are compatible with infinite-size grammatical expressions ... |

284 | Decidability of second-order theories and automata on infinite trees - Rabin |

213 | Weak Second-order Arithmetic and Finite Automata - Büchi - 1960 |

169 | Tree acceptors and some of their applications - Doner - 1970 |

161 | Finitary models of language users - Miller, Chomsky - 1963 |

138 | On a decision method in restricted second-order arithmetics - BÜCHI - 1962 |

126 | Automata and Computability - Kozen - 1997 |

117 |
Recursively enumerable sets of positive integers and their decision problems
- Post
- 1944
(Show Context)
Citation Context ...gs of words corresponding to natural language sentences, did not come from any classical or traditional source; it was original with Chomsky, who had read Post and acknowledges the intellectual debt (=-=Post 1944-=- was Chomsky’s cited source for the term ‘generate’; see Chomsky 1959:137n). Today, virtually all theoretical linguistics that aims at explicitness is based on Chomsky’s development of Post. Chomsky’s... |

105 | Finite automata, formal logic, and circuit complexity - STRAUBING - 1994 |

99 | Derivational minimalism
- Stabler
- 1997
(Show Context)
Citation Context ... socalled ‘minimalist’ grammars of Chomsky’s recent transformationalist work appear to be very similar to categorial systems in this regard, and if Edward Stabler’s formalization of them is accepted (=-=Stabler 1997-=-) they are equivalent to multi-component tree-adjoining grammars (Harkema 2001; Michaelis 2001).sContrasting Applications of Logic in Natural Language Syntactic Description 5 Mathematical sources of m... |

82 |
Formal Reduction of the General Combinatorial Decision Problem
- Post
- 1943
(Show Context)
Citation Context ...Post’s goal was a mathematicization of the notion of a proof in logic. In pursuit of this he developed a schema defining the general form of inference rules in terms of what he calls productions (see =-=Post 1943-=-:197). A production has a finite number of ‘premises’ and produces one ‘conclusion’. Each of these is a finite string of symbols. A production thus maps a set of strings to a string, has the form seen... |

75 |
Recursive unsolvability of a problem of Thue
- Post
(Show Context)
Citation Context ...symbol. The Type 0 rules of Chomsky (1959)) are defined by a restriction very similar in spirit. Type 0 grammars are essentially Post’s semi-Thue systems, deriving from earlier work by Axel Thue (see =-=Post 1947-=-). Again, n in (1a) is limited to 2 so that productions have the form σ1 ⇒ σ2, but: – m in (1b) is limited to 2 (i.e., there are just two variables in each production), so σ1 and σ2 both have the form... |

59 |
On Generative Semantics
- Lakoff
- 1971
(Show Context)
Citation Context ... by context-sensitive rules under the tree-admitting interpretation is the set of context-free stringsets. A wider application of the same kind of thinking about how to state grammars can be seen in (=-=Lakoff 1971-=-). Lakoff proposes that the structure of a natural language expression should be idealized as a finite sequence of finite trees 〈∆0, . . .,∆k〉 in which ∆0 is by definition the deep structure of the ex... |

58 |
Indexed grammars - an extension of context-free grammars
- Aho
- 1968
(Show Context)
Citation Context ...be characterized in terms of wmsol on three-dimensional tree manifolds — roughly, trees that have trees as their nodes; and Langholm (2001) has shown that the much larger class of indexed stringsets (=-=Aho 1968-=-) can also be characterized with a particular kind of bounded existential quantification in wmsol on trees with added links between the nodes. All the work just reviewed uses model theory to character... |

52 | A Descriptive Approach to Language-Theoretic Complexity
- Rogers
- 1998
(Show Context)
Citation Context ...gs in the work of Büchi (1960). 3 Büchi’s results were motivated by questions of arithmetic. It took more than thirty years for their relevance to linguistics to be appreciated (e.g., by Rogers 1994, =-=Rogers 1998-=-, and Kracht 2001). The basic question concerned the expressive power of logic of a particular kind for talking about finite sequences. The logic in question was ordinary predicate logic augmented in ... |

45 | Talking about trees
- Blackburn, Gardent, et al.
- 1993
(Show Context)
Citation Context ...an to appear. First, Patrick Blackburn and his colleagues, and independently Marcus Kracht, began to develop an idea from Gazdar et al. (1988): using modal logic as a description language for syntax (=-=Blackburn et al. 1993-=-; Kracht 1993; Blackburn and Gardent 1995; Blackburn and Meyer-Viol 1997). And second, James Rogers began his work on applying wmsol to linguistics (Rogers 1994, 1996, 1997, 1998, 1999). These papers ... |

45 | The MIT Encyclopedia of the Cognitive Science - Wilson, Keil - 1999 |

34 | Mutual beliefs in question-answer systems - Joshi - 1982 |

34 | Syntactic codes and grammar refinement
- Kracht
- 1995
(Show Context)
Citation Context ...trick Blackburn and his colleagues, and independently Marcus Kracht, began to develop an idea from Gazdar et al. (1988): using modal logic as a description language for syntax (Blackburn et al. 1993; =-=Kracht 1993-=-; Blackburn and Gardent 1995; Blackburn and Meyer-Viol 1997). And second, James Rogers began his work on applying wmsol to linguistics (Rogers 1994, 1996, 1997, 1998, 1999). These papers from the 1990... |

33 | On the distinction between model-theoretic and generative-enumerative syntactic frameworks - Pullum, Scholz - 2001 |

29 |
Studies in the Logic of Trees with Applications to Grammar Tbrmalisms
- Rogers
- 1994
(Show Context)
Citation Context ...ical beginnings in the work of Büchi (1960). 3 Büchi’s results were motivated by questions of arithmetic. It took more than thirty years for their relevance to linguistics to be appreciated (e.g., by =-=Rogers 1994-=-, Rogers 1998, and Kracht 2001). The basic question concerned the expressive power of logic of a particular kind for talking about finite sequences. The logic in question was ordinary predicate logic ... |

29 | Category Structures - Gazdar, Pullum, et al. - 1988 |

28 |
Concerning the base component of a transformational grammar. Foundations of Language
- McCawley
- 1968
(Show Context)
Citation Context ...nt of the assignment of degrees of ungrammaticality. Degree of ungrammaticality depends entirely on lexical category sequence similarities, according to this proposal; 9 Incidentally, we would follow =-=McCawley 1968-=- in taking the deviance of selectional restriction violations like Golf plays John as semantic rather than syntactic.s12 Geoffrey K. Pullum and Barbara C. Scholz it does not depend on any aspect of sy... |

26 | Parsing minimalist languages - Harkema - 2001 |

21 |
The openness of natural languages
- Postal
- 2004
(Show Context)
Citation Context ...e premises would be false, while if English were infinite the argument would be circular, and we are given no way to tell which is the case. (This point is not new; we are basically just paraphrasing =-=Langendoen and Postal 1984-=-:30-35.)sContrasting Applications of Logic in Natural Language Syntactic Description 17 If some argument did show that English or some other natural language had infinitely many expressions, the fit w... |

21 | Context sensitive immediate constituent analysiscontext-free languages revisited - Peters, Ritchie - 1969 |

19 | Model theory and the content of OT constraints - Potts, Pullum - 2002 |

17 | Transforming linear context-free rewriting systems into minimalist grammars. In: de Groote et al - Michaelis |

14 |
Explanatory Models in Linguistics
- Chomsky
- 1962
(Show Context)
Citation Context ... infinite number of strings representing grammatical sentences and a ‘morphophonemic component’ that specifies the physical shape of each of these sentences. This is the classical model for grammar. (=-=Chomsky 1962-=-:539) But such a model is not “classical” in any sense. The idea that grammars of natural languages generate infinite sets of strings was only a few years old in 1960, and the mathematics supporting i... |

13 | Syntax: A Generative Introduction - Carnie |

12 | A model-theoretic framework for theories of syntax - Rogers - 1996 |

12 | Grammarless" phrase structure grammar
- Rogers
- 1997
(Show Context)
Citation Context ...hat the daughter labels are but also whether the unique B is contained within the subtree dominated by the current node.sContrasting Applications of Logic in Natural Language Syntactic Description 7 (=-=Rogers 1997-=-:739). Then the notion of being privileged with respect to feature f can be defined by Privileged f ≡ ∀X[P ′ f(X) → X(x)] (Rogers 1997:740). Feature specification defaults become easily statable as si... |

11 |
Modal logic and model-theoretic syntax
- Blackburn, Meyer-Viol
- 1997
(Show Context)
Citation Context ...ndependently Marcus Kracht, began to develop an idea from Gazdar et al. (1988): using modal logic as a description language for syntax (Blackburn et al. 1993; Kracht 1993; Blackburn and Gardent 1995; =-=Blackburn and Meyer-Viol 1997-=-). And second, James Rogers began his work on applying wmsol to linguistics (Rogers 1994, 1996, 1997, 1998, 1999). These papers from the 1990s, however, were largely preoccupied with restating particu... |

11 | Abstract syntax - Keenan, Stabler - 1994 |

10 | A specification language for lexical functional grammars
- Blackburn, Gardent
- 1995
(Show Context)
Citation Context ...rn and his colleagues, and independently Marcus Kracht, began to develop an idea from Gazdar et al. (1988): using modal logic as a description language for syntax (Blackburn et al. 1993; Kracht 1993; =-=Blackburn and Gardent 1995-=-; Blackburn and Meyer-Viol 1997). And second, James Rogers began his work on applying wmsol to linguistics (Rogers 1994, 1996, 1997, 1998, 1999). These papers from the 1990s, however, were largely pre... |

10 | Some methodological remarks on generative grammar - Chomsky - 1961 |

9 | The logical structure of linguistic theory. Unpublished manuscript. Published as a book with a new introduction in 1975 by - Chomsky - 1955 |

8 | Decision problems of finite automata and related arithmetics - Elgot - 1961 |

8 | A regular characterization of graph languages definable in monadic second-order logic, Theoret - Engelfriet - 1991 |

8 |
Scandinavian extraction phenomena revisted. Weak and strong generative capacity
- Miller
- 1991
(Show Context)
Citation Context ...t all. There are a few central clausal constructions in other languages that go beyond the recognition power of standard tree automata, but they have so far been found only in Germanic (Shieber 1985; =-=Miller 1991-=-), and they seem quite rare. Work in computational linguistics has shown that if we have the power to describe the recognizable treesets, we have enough descriptive power for doing most of the work ne... |

8 | 2003: wMSO theories as grammar formalisms - Rogers |

6 |
On certain formal properties of grammars. Information and Control 2(2):137–167. Reprinted
- Chomsky
- 1959
(Show Context)
Citation Context ...texts where X1 precedes and X2 follows. Again there is no loss of expressive power from Post’s unrestricted production systems over vocabularies of terminal and nonterminal symbols (this is proved in =-=Chomsky 1959-=-): all (and only) the r.e. sets over the terminal vocabulary can be generated. The idea of cutting productions completely adrift from the formalization of inference calculi, and using them instead to ... |

6 | The Descriptive Complexity of Generalized Local Sets - Rogers - 1999 |

5 |
Evidence against the context-freeness of human language
- Shieber
- 1985
(Show Context)
Citation Context ...se structure at all. There are a few central clausal constructions in other languages that go beyond the recognition power of standard tree automata, but they have so far been found only in Germanic (=-=Shieber 1985-=-; Miller 1991), and they seem quite rare. Work in computational linguistics has shown that if we have the power to describe the recognizable treesets, we have enough descriptive power for doing most o... |

5 | Bare grammar - Keenan, Stabler - 2004 |

4 | A descriptive characterisation of indexed grammars - Langholm - 2001 |

4 | The best theory - Postal - 1972 |

3 | Rule orderings, obligatory transformations, and derivational constraints. Theoretical Linguistics - Soames - 1974 |

2 |
Syntactic theory, Holt Rinehart and
- Bach
- 1974
(Show Context)
Citation Context ... concoct some way to add at least one word to the sentence and come up with a longer English sentence, then we are driven to the conclusion that the set of English sentences is (countably) infinite. (=-=Bach 1974-=-:24) b. Is there anyone who would seriously suggest that there is a number, n, such that n is a number is a sentence of English and n+1 is a number is not a sentence of English... ? On this basis we t... |