FREE ELECTRONIC LIBRARY - Abstracts, online materials

Pages:     | 1 || 3 | 4 |   ...   | 8 |

«Formal Phonology* David Odden OSU Abstract Two problematic trends have dominated modern phonological theorizing: over-reliance on machinery of ...»

-- [ Page 2 ] --
cost-free. Formal Phonology does not start with a body of a priori technical claims about grammars. FP is a metatheoretical approach which defines the nature of the enterprise, and it embodies an epistemology which allows valid hypotheses to be advanced regarding the nature of phonology. Such substantive claims as are made herein are claimed to be those which are justified by what is known. Thus the conclusion that phonological rules are formalize with featural expressions (not lists of atomic phonemes) would follow from the methodology of FP and the facts of human language, but would not be a foundational stipulation of the theory. The next section discusses the scope of inquiry in phonology, and the nature of inductive epistemology.

3. Domain restriction and falsification: the significance of modularity Before delving extensively into the logic of theorizing in FP, it is necessary to be explicit about what the object of study is – it is futile to ask questions about the methodology of investigating an object, without first identifying what the object is. FP claims that an autonomous phonological component in grammar is necessary in order to understand linguistic behavior. FP holds that the concern of theoretical phonology is modeling the form of computations in the phonological component, including the things that computations are performed on. The kind of fact that FP is about is that Kimatuumbi /mu-wɪkili-ɛ/ surfaces as [ŋŋwɪkɪlí] ‘2pl should cover’, Classical Arabic /tawaḍiʕu/ becomes [taḍaʕu] ‘she lays’, and Karanga Shona /mu-á-ka-mú-bikira/ is realized as [makámubíkira] ‘2pl cooked for him’. A study of “language sound” would include not just the study of grammars, but also a lot of speech behavior that is not about grammar.

No theory of “language sound” can be correct without a theory of phonological grammars.

As observed in Hale (2007), Hale & Reiss (2008) and elsewhere, production and perception of English “cat” involves more than the grammatical representation and computation of the output [khæt] from the input /kæt/. The entire chain of events involved in speech transmission or reception involves many non-grammatical and non-linguistic, indeed non-biological factors, which are outside the concern of phonological theory. The grammatical mapping from /kæt/ to [khæt] is squarely in the domain of a theory of phonology. FP is “Galilean”, in abstracting away from matters of performance (speaker attention, error, etc.) or acoustic differences between “cat” uttered on a cold, dry day versus a hot, humid day. FP is also concerned only with the grammatical properties of the phonological component, meaning that a formal phonology is not also responsible for accounting for what the syntactic, morphological, semantic or phonetic components of a grammar do.

Phonology being just one aspect of a general theory of language behavior (which is itself one aspect of cognition), not all facts about language sound are in the explanatory domain of phonological theory.

Even facts which are “about language” and not environment may be outside the domain of a formal theory of phonology. For example, some languages have a process of post-nasal voicing (Kimatuumbi), and others have a process of post-nasal devoicing (Setswana). It is the responsibility of a formal theory of phonology to enable the description of both states of affairs, since both exist. Post-nasal devoicing is extremely rare (being found in only one group of Bantu languages), but post-nasal voicing is relatively well attested, being found in many Bantu languages as well as Greek, Japanese, Zoque, Maasai, and Imbabura Quechua. A formal theory of phonology is not held responsible for encoding this statistical generalization, since there already exist substantive theories of linguistic behavior including language acquisition, physiology and physics which account for this distributional asymmetry (see Hayes & Stivers 1995), and therefore the observation that a certain fact is “rare” or “marked” is irrelevant. FP only asks what the form of a E.g. Yip (1988: 76) “The outside trigger is of course the OCP, a universal principle and thus free of charge”.

The converse is not true. The reason for the asymmetry is the simple fact that “language sound systems” properly includes “phonology”, and not vice-versa.

This assumes that the output of the phonology is in fact [khæt], which may not actually be the case.


phonological computation is, not what its likelihood is. Even a zero practical probability of occurrence does not put an observational gap in the domain of grammatical theory.

The potential correctness of a theory of phonological grammars is also not impugned by observing that the elements of the theory can combine in a way that correspond to so-far unattested sound-pattern phenomena. To restate the point, FP only asks what the form of a phonological computation is, not what it isn’t. The situation where a theory can describe patterns that have not been observed is often seen as evidence that the theory is “too powerful”, that it “overgenerates”. For instance, numerous formal theories of phonology would allow the description of the hypothetical rule /p b f v m/ → [t d s z n]/__[y]. Such a

process is not yet known to exist in any language. The fact that we have not encountered this event is almost certainly outside the purvue of FP, since the process is formalizable in any general, empiricallyadequate theory of phonology. It is obviously expressible as a feature-changing rule:

(1) [+ant,+cons] → [+cor] / __ [+hi,-back,-syl] It is also expressible as cross-planar spreading in a multilinear representational theory where front vowels and glides are [+coronal] – see Hume (1994) for the treatment of analogous changes of velars, triggered by front vowels and glides. Such a rule would not be expressible in a theory where all rules must be expressed as deletions under featural identity or as spreading if [y] is [-coronal], but such a theory is empirically untenable, since attested rules of the type /k g x/ → [č ǰ š] / __ [y] would then also be unformalizable. The fact that labial-coronalization is unattested is, then, not the result of the phenomenon being intrinsically unformalizable, therefore the gap is outside the scope of what a theory of phonology must explain. (Rice 2007 discusses a distinct sense of “gap” which is squarely in the purvue of grammatical theory, where affixation is unexpectedly blocked in a defined environment, e.g. in Norwegian where imperatives do not exist for verbs whose roots end in unsyllabifiable clusters).

It might seem to be within the scope of FP to explain why a rule /p z ŋ/ → [γ ɬ t] / __ [y] is unattested. The reasoning could be that the collection of consonants /p z ŋ/ cannot be described using known tools for referring to subsets of a segmental inventory, and no phonetic property can be abstracted to describe the mappings {p→γ, z→ɬ, ŋ→t}. This argument logically depends on two premises – ones which must be previously established. The first, which is valid and indispensible in Formal Phonology and can be considered to have been established, is that segment classes and phonological changes are defined via conjunctions of features. It also requires features to be defined in terms of substantive universals so that the set {p,z,ŋ} and the respective structural changes could not be formally expressed. The latter is a questionable claim, not necessarily accepted in FP (it is a possible claim under FP), requiring justification.

A formal theory of phonological computation cannot be refuted by phenomena, and phonological phenomena are computationally epiphenomenal. A theory of phonological computations is refutable only by facts about phonological computations that contradict the theory. Phonological phenomena become relevant to a theory of computation only when there is a compelling pairing of a theory of representations and a theory of computation which renders the grammatical description of an attested phenomenon impossible, given those theories (the concern being, of course, that a theory which cannot describe actual facts is wrong, qua theory of language).

It is the proper concern of a Formal Phonological account of a specific language to say whether there is a rule of intervocalic voicing, or post-vocalic spirantization, or final devoicing in that language. If underlying /apa/ maps to surface [aba], then the phonological grammar must contain a rule or similar formal object which performs that mapping. It is also the proper concern of a Formal Phonology metatheory of grammar to determine whether phonological mappings involve string-changing mappings or stringfiltering constraints. Therefore, this question must be decided empirically given the metatheoretical reOne could imagine that the lack of a rule /py/ → [t] could be predicted by a well-motivated theory of phonological computation (though nothing presently known suggests that any such theory is possible). But such a prediction would be

an accident. The theory would not be founded on the desideratum of formally precluding a rule of labial coronalization:

instead, the principles which hypothetically yield this result would be independently justified on the basis of other facts.


quirements of FP, and not stipulated arbitrarily as a theoretical postulate. Other valid concerns of Formal Phonology are whether the operation embodied in a rule applies to just a single segment or can simultaneously apply to multiple (perhaps unbounded) segments; whether rule or constraint statements include universal and existential quantifiers or just universal quantifiers; whether references to substrings identified by such statements involve conjunction and disjunction or just conjunction (these questions must be decided empirically given the metatheoretical requirements of FP, and not stipulated arbitrarily as a theoretical postulate). These are matters about the form of rules, which is the concern of Formal Phonology.

The fact that rules of intervocalic devoicing or post-stop spirantization have yet to be uncovered in grammars does not justify adding new theoretical concepts to prohibit such rules, since the nonexistence of such rules is already explained via theories of learning and historical change, and duplicating functional reasons for the non-existence of intervocalic devoicing in the computational apparatus would be otiose.

The lack of examples of intervocalic devoicing could imaginarily be “explained” in grammar by positing some complex of added notions about feature changing, the context “between vowels”, and values of voicing; but such a complication would be inferior to the simple phonology-external fact that intervocalic devoicing requires unlikely phonetic mechanisms to bring it about. Adding formal principles to phonology to say that intervocalic devoicing is not a computationally-possible rule contributes nothing, since there is nothing in the form of the computation that is “impossible”. It is thus a basic principle of Formal Phonology that the lack of instances of a certain kind of rule does not compel complication of the computational theory.

Likewise, grammatical reification of non-phonological explanations for the existence of common rules, in the form of grammatical “benefits” for rules like intervocalic voicing or g-spirantization, does not contribute anything to our understanding of grammars. It is known that across languages, the voiced velar obstruent stop [g] has a greater tendency to change to something else – voiceless, fricative, or sonorant – than do labial or alveolar voiced stops. The explanation for this derives from non-linguistic facts about airflow, human anatomy, and the physiology of vocal fold vibration. Since the explanation for the propensity of /g/ to change already has an explanation (Boyle’s Law is an independent fact of physics, Bernoulli’s Principle is an independent fact of physics, the location of the constriction in a velar is a independent fact of articulation, the mass of the vocal folds is an independent fact of anatomy), re-stating the sum of these factors as an autonomous principle of grammatical computations is entirely redundant, adding nothing to our knowledge of the universe. See Hale & Reiss (2000, 2008) for extended discussion of the logical problems with duplicating principles of phonetics and learning within grammatical theory.

We return to the question of post-hoc “functional rationalization” of grammatical principles in section 5.

It should be clear from the preceding discussion that considerations of overgeneration in the language-enumerating sense play a minor role in theory-evaluation in Formal Phonology (see the next section for discussion of the proper role of overgeneration concerns in theory construction). FP does not thereby open up the theoretical floodgates and say “Everything is formally possible; the explanation for all unattested patterns lies in functional factors”. Such a move would be equivalent to denying the existence of phonology, which FP does not do. FP does, however, deny that phonological theory shoulders the sole explanatory responsibility for the facts of speech behavior. It is fair to say, though, that Formal Phonology is not particularly concerned over the fact that a theory allows unattested language “types”, when the required theoretical devices are well-justified. The concern of FP is, instead, over what theoretical devices are required to describe the nature of phonological grammars.

This is not to imply that phonetics is entirely irrelevant to phonology. First, it is relevant on practical grounds because a theorist has to know if a generalization about language has an independent, non-phonological explanation. Second, a theory of grammar must ultimately mesh with a theory of physical implementation as part of a grander theory of the mind, and if some theory of phonology patently contradicts what is clearly true about physical implementation, then that theory of phonology cannot be correct. It is crucial, though, that the theory of physical implementation be “clearly true”, not just “somewhat supported” or “the current belief”.


Pages:     | 1 || 3 | 4 |   ...   | 8 |

Similar works:

«QEH Working Paper Series – QEHWPS140 Page 1 Working Paper Number 140 From ‘OPEN SEASON’ to ‘ROYAL GAME’: The Strategic Repositioning of Commercial Farmers across the Independence Transition in Zimbabwe 1972-1985 Angus Selby1 This paper explores the strategic repositioning of commercial farmers across the Independence transition, from a close proximity to the Rhodesian Front to an alliance with the Mugabe regime. It argues, contrary to most analyses, that commercial farmers were...»

«FOLLOWING THE VOLGA VIKINGS John Vallentine (The author, a retired doctor, left his native Australia in 2006 and is halfway through ‘the world’s slowest circumnavigation’ – a contrast to his previous Sydney-Hobart racing days in ‘a fast 1 Tonner’. In his mid sixties, he admits to ‘feeling much older during long night watches in bad weather’. Tainui is his fifth offshore yacht and, he says, he ‘cannot imagine a better cruising vessel.’ A 1973-built Peterson 46, she is...»

«Precast concrete blocks used for highway emergency parking lane pavement engineering practice Wang Huoming 1 Nie Xin1 Zhou Gang1 Cai Congbing 2 Xiong Shiyin 3 (1 China Merchants Chongqing communications research and Design Institute Company Limited Chong Qing 400067; 2 Yunnan Luqiao Limited by Share Ltd Kun Ming 650200; 3 Mo Si Expressway Development Co Pu 'er 665100 ) Abstract: Prefabricated concrete pavement is composed of a single precast block assembled as a special pavement structure...»

«EQUAL: PAVING THE WAY FOR LIFELONG LEARNING AND AGE MANAGEMENT Published in November 2007 The contents of this publication were prepared by GHK Consulting Ltd for the European Commission and do not necessarily reflect the opinion or position of the European Commission, Directorate-General for Employment, Social Affairs and Equal Opportunities. CONTENTS Executive Summary 1. Introduction: Why EQUAL? 2. Engaging employers – especially small and medium-sized enterprises 2.1. Practical solutions...»

«  Creditor Interventions and Firm Innovation: Evidence from Debt Covenant Violations* Yuqi Gu Connie X. Mao Xuan Tian Pennsylvania State University Temple University Indiana University yug16@psu.edu  cmao@temple.edu tianx@indiana.edu (610) 725-5391 (215) 204-4895 (812) 855-3420     This version: March, 2014                                                              * We remain responsible for any remaining errors or omissions.  ...»

«BOOK 2 OF 3 A DETAILED LOOK at Projects and Plans W E ’ R E L I V I N G O U T O U R M I S S I O N S TAT E M E N T : To see every[one] transformed and mobilized by Jesus We’re putting it all out there. We’re giving you the blueprints to our bold plans for the next chapter of Kensington if we reach our goal of $30M. In the following pages, you’ll read about the dreams for our new hub facilities, how we’re recharging our existing hubs, our commitment to growing strong campuses, our...»


«ActA Univ. SApientiAe, SociAl AnAlySiS, 6, 1 (2016) 71–91 How Professionals in the Social System and Child Protection Perceive Their Profession Eszter PAPP University of Debrecen, Hungary e-mail: pappeszter93@gmail.com Andrea RÁCZ University of Debrecen, Hungary e-mail: raczrubeus@gmail.com Abstract. Our research is based on the analysis of semi-structured interviews made with 20 professionals in social care and child protection. The research aims at finding the answers to how the legacy...»

«1 NEMAAR Conference Program April 26, 2014 Boston College School of Theology & Ministry (STM), 9 Lake St., Brighton MA 02135 Handicapped parking and access to the STM building is available from the rear (east) entrance at the ground level. Restrooms and elevators are located at the far north end of each floor. Vending machines are on the lower level mid-corridor. 8:15—8:55 AM Registration in the BC-STM Foyer of 9 Lake St. 9:00—10:15 AM First Concurrent Session of Papers Room 100 (north end)...»

«Using a Foundational Ontology for Reengineering a Software Enterprise Ontology Monalessa Perini Barcellos1,2 and Ricardo de Almeida Falbo1 Ontology and Conceptual Modeling Research Group (NEMO), Federal University of Espírito Santo, Brazil COPPE, Federal University of Rio de Janeiro, Brazil {monalessa, falbo}@inf.ufes.br Abstract. The knowledge about software organizations is considerably relevant to software engineers. The use of a common vocabulary for representing the useful knowledge about...»

«1 CAISTOR AREA DIARY NEW ENTRY & EVENT LISTING weekending FRIDAY NOVEMBER 21st compiled for the media & community as a free promotional service for the wider Caistor Area. Alan Caine 01472 851674 alncn100@gmail.com Please note new online REGULAR EVENTS calendar http://www.localendar.com/public/alncn100 CAISTOR MARKET EVERY WEDNESDAY & SATURDAY NEW ENTRY SECTION: WEEK ENDING FRIDAY NOVEMBER 21st 2014 For Media Diaries etc, these are repeated in Chronological diary below I have no crystal ball so...»

«Offices across Asia, Africa and Latin America www.MicroSave.net info@MicroSave.net ASA’s Culture, Competition and Choice: Introducing Savings Services into a MicroCredit Institution Graham A.N. Wright, Robert Peck Christen and Imran Matin July 2001 MicroSave – Market-led solutions for financial services ASA’s Culture, Competition and Choice: Introducing Savings Services into a MicroCredit Institution – Wright et al. 1 Table of contents Executive Summary Introduction 4 ASA – A...»

<<  HOME   |    CONTACTS
2017 www.abstract.dislib.info - Abstracts, online materials

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.