Date: 04-Aug-2006
From: Robert Felty <robfeltyumich.edu>
Subject: Cross-Linguistic Aspects of Processability Theory
Announced at http://linguistlist.org/issues/17/17-392.html
EDITOR: Pienemann, ManfredTITLE: Cross-Linguistic Aspects of Processability TheorySERIES: Studies in Bilingualism 30PUBLISHER: John BenjaminsYEAR: 2006
Robert Felty, Departments of Linguistics and German, University of Michigan
This book is in essence a continuation and expansion of ProcessabilityTheory (PT) as proposed by Pienemann (1998). In that book he laid out atheory of language acquisition based on psycholinguistic principles, andapplied it to data from several languages (English, German, Japanese, andSwedish). In this edited volume, Pienemann and five other scholars furtherapply PT to several more typologically diverse languages (Arabic, Chinese,Japanese), which results in an extension of the theory.The book is divided into eight chapters. The first chapter provides asummary of Pienneman's (1998) original proposal of PT. Chapters twothrough six contain applications of PT to various languages. In chapterseven, Pienemann proposes an extension to PT based on the precedingevidence. Further empirical evidence is discussed in chapter eight, usingthe extended theory of PT.
SUMMARY
In the first chapter, Pienemann outlines the basic tenets of PT and severaltests of its applicability to real language data. PT is heavily based ontwo other theories: Lexical Functional Grammar and the model of languageproduction proposed by Levelt (1989). His theory differs in three keyaspects from many other theories of language acquisition: (1) processing isboth incremental and parallel, (2) language acquisition is constrained byhuman psychological constraints, and (3) first language (L1) and secondlanguage (L2) acquisition occur by the same procedures. The first aspectis drawn largely from Levelt (1989), and is used as a way to explain theso-called ''logical problem'', i.e. how do children acquire language soquickly with such limited input. If language processing is not equivalentto one CPU trying to parse an entire sentence, but rather has severaldifferent processors working in parallel, this greatly increases theefficiency of learning language. The second aspect is a response to theprevalent views of learnability theory, which also address the logicalproblem of language acquisition, but merely from a computationalperspective. The third aspect is in direct contrast with a very prevalentview in second language acquisition (SLA) which state that L1 acquisitionis achieved by access to universal grammar (UG), whereas access to UG iseither limited or non-existent during L2 acquisition, and thus otherstrategies must be used.
After laying out the basic aspects of the theory, Pienemann briefly appliesthe theory to data from previous studies on SLA, including several onEnglish and German. He then goes on to compare L1 and L2 acquisition ofGerman word order, and shows that the different developmental paths canboth be accounted for by PT, by means of the notion of ''generativeentrenchment'', a notion he adopts from biology. The underlying principleof generative entrenchment is that development occurs in small steps, andthat earlier choices constrain later choices. In this manner, an incorrectchoice early on forces the path of L2 development down a very differentroute than L1 development, though they both (potentially) end up at thesame state.
The second chapter consists of a brief history of the development of PTprior to the 1998 proposal, as well as several developments thereafter.Pienemann and colleagues have proposed several different theories andmodels aimed at describing and explaining SLA beginning in the early 1980s. The first of these was the Multidimensional Model (Meisel, Clahsen, &Pienemann 1981), which was primarily a descriptive framework for dynamicprocesses in L2 development. The second theory discussed here is theStrategies Approach (Clahsen 1984), which proposes an explanation forGerman L2 word order development. Pienemann explains that PT incorporatesmost of the successes of these theories, as well as aspects of theTeachability Hypothesis (Pienemann 1984, 1989) and the Predictive Framework(Pienemann & Johnston 1987), while addressing the shortcomings of thesetheories.
The third chapter marks the beginning of the meat of the book -- newstudies on SLA from a PT perspective. In 'Processability, typologicaldistance, and L1 transfer', Pienemann, Biase, Kawaguchi and Håkanssoncompare the predictions of various SLA theories with regards to L1transfer. Many theories of SLA take the final state of the L1 grammar asthe starting point for L2 acquisition, and therefore predict that anyrelevant aspects of the L1 should be transferred to the L2. PT takes avery different approach however, namely that the starting point for L2acquisition is virtually a blank slate, and that any components of the L1that could be transferred will only do so at the appropriate stage in theacquisition hierarchy. The authors cite some fairly convincing evidence insupport of this view. The data from bilinguals and learners paired bylanguage (e.g. L1 Finnish, L2 Swedish and vice versa) are particularlycompelling. These data clearly show that even when the learners L1 and L2share a feature (e.g. Subject Verb inversion after initial adverbs inGerman and Swedish), the learners might show a different pattern, dependingon their level of acquisition, as based on PT's acquisition hierarchy.Moreover, it is also shown that the degree of L1 transfer is notbi-directional in these cases.
The first language specific paper is provided in chapter four by FethiMansouri entitled 'Agreement morphology in Arabic as a second language'.The primary contribution from this chapter is the addition of evidence insupport of PT from a language typologically very different from previousexamples. The first part of this chapter is devoted to describing aspectsof Arabic morphology and syntax in an LFG framework, and using this todevelop a proposed acquisition hierarchy. This hierarchy is then supportedby evidence from Mansouri's (2000) year-long longitudinal study of twoAustralians acquiring Arabic in a classroom setting.
Similar to chapter 4, chapter 5, 'Processing and formal instruction in theL2 acquisition of five Chinese grammatical morphemes', by Yanyin Zhang isdivided into two parts: an LFG account of the morphemes in question, and anSLA study of three Australians learning Chinese in a classroom setting. Inthe first part an LFG analysis of Chinese morphology is proposed, and anacquisition hierarchy is constructed. The empirical evidence support a PTaccount of Chinese L2 acquisition, adding to the universal claim of PT.Zhang also analyses the impact of formal instruction on acquisition. Foreach grammatical structure analyzed, emergence was found only after thestructure had been presented in class, though some learners acquired itmuch sooner than others. Formal instruction did not have an impact on theorder of the acquisition hierarchy however -- that is, formal instructioncan aid in speeding up the process of acquisition, but all the stages ofthe hierarchy must be followed in order.
One of the strengths of PT is displayed in chapter 6, 'Similarities anddifferences in L1 and L2 development'. In this paper, Håkansson firstcompares previous findings of L1, L2 and SLI (Specific Language Impaired)learners. She then presents her own findings from a longitudinal study withSwedish learners, with five learners from each group. As mentionedearlier, PT can be applied both to L1 and L2 acquisition, and in this paperthe advantage of this capability is born out. Håkansson is able to use thesame acquisition hierarchy to discuss all three groups of learners, andfinds three different patterns of acquisition, though the SLI learners seemto pattern more closely to L2 learners than L1. She uses this finding tosuggest that perhaps acquisition should not be viewed as simply a dichotomybetween L1 and L2, but rather a continuum.
In chapter seven, 'Extending Processability Theory', Pienemann, Di Biaseand Kawaguchi incorporate new aspects of LFG into PT in order to betterexplain a wider range of data. Two components are added to PT in thischapter: (1) The Lexical Mapping Hypothesis, and (2) The Topic Hypothesis. The former is a relatively new feature of LFG (Bresnan 2001), whichexplains mappings from constituent structure (NP, VP, PP, etc.) tofunctional structure (FOCUS, SUBJ, OBJ, etc.). The authors hypothesizethat there is a canonical mapping, in which the order of the c-structuredirectly follows the f-structure, and that operations such as subject-verbinversion and question formation deviate from this canonical mapping, andare therefore difficult for L2 learners to process. The topic hypothesisclaims that L2 learners will not initially differentiate between SUBJ andTOPIC, which will cause a different pattern of acquisition than L1learners. Like many of the other chapters of the book, the first partfocuses on formalisms in LFG, and the second part focuses on itsapplication to SLA. Though this chapter was fairly dense with theory,numerous examples aided in comprehension.
The extensions proposed in chapter seven are utilized by Kawaguchi in theeighth chapter, 'Argument structure and syntactic development', in whichthe L2 acquisition of Japanese syntax is investigated. Before laying outan LFG account of Japanese syntax, Kawaguchi briefly reviews previousinvestigations of Japanese acquisition. After an acquisition hierarchy isconstructed, empirical data from two longitudinal studies conducted by theauthor, one two-year and one with three-year study, with one learner each. In both studies, the proposed acquisition hierarchy is supported. Thecontribution of this chapter lies in the use of the newly incorporatedaspects of PT, without which the analysis would not have possible.
EVALUATION
Both in the present volume and in the 1998 book, Pienemann makes some boldclaims as to the nature of language acquisition, and backs these claims upwith substantial data. The data is the main focus of this volume, whichprovides much further evidence in support of PT. More importantly though,some of the data also show limitations of the theory as originallyproposed. Pienemann, Biase, and Kawaguchi are to be commended for proposingan extension to the theory, rather than simply throwing it out in favor ofa different theory. A remaining question is whether PT can also account foracquisition of phonology and phonetics, and if not, whether an additionalextension to the theory is possible. Phonetics might be particularlydifficult, as phonetic data is inherently continuous, whereas PT is verydiscrete in nature.
At times Pienemann comes across as somewhat defensive, especially in thesecond chapter, in which he dedicates 6 pages to rebuffing criticism of PTposed by Jordan (2004). While he goes into a fair amount of detail aboutthe criticisms, it seems slightly out of place here. Surely to the readerencountering PT for the first time, this discussion is confusing at best,since it cannot fully be understood without having first read bothPienemann (1998) and Jordan (2004). In particular, the notion of the'emergence criterion' is discussed in length, but once again, for adefinition thereof, the reader is simply referred back to the 1998 volume. Since this is a key point in understanding PT, it would have beenextremely helpful to provide a definition of the concept in the present work.
One significant drawback for PT is that it relies on a very specific typeof data -- production data in a natural setting from a modest number ofsecond language learners. In order to be able to construct a hierarchy ofstages of acquisition, a large amount of data must be collected. Thenevidence for or against the acquisition of particular structures must befound at each stage in the hierarchy in order to confirm or falsify thepredictions of PT. Some might argue that this is the only research methodthat is truly valid, but unfortunately the limits of time and money in thereal world often prevent researchers from undertaking such large projects.The other possible drawback with using longitudinal data is that frequentlythe number of participants is very small (including several studies in thisvolume), and thus the generalizability of the study is questionable.Moreover, others might argue that this approach ignores data from othertasks such as speech perception and grammaticality judgments, which canalso provide information valuable information about stages of acquisition.
Overall the editorial quality of the book is quite high. Besides theoccasional typographical error, there were only two major errors to befound: (1) Two tables on page 148 that are intended to show differences inthe acquisition of two learners are in fact the same table, which yieldsthe claims of difference unverifiable (2) the last chapter contains anumber of grammatical errors which should have been corrected by the editor.
In spite of the few shortcomings mentioned above, 'Cross-LinguisticsAspects of Processability Theory' should prove a valuable read for anyoneinterested in language acquisition, particularly SLA, and most specificallywith the acquisition of morphology and syntax. Pienemann and colleagueshave presented very thought-provoking ideas for linguists and psychologistsinterested in language acquisition to ponder.
REFERENCES
Bresnan, Joan. 2001. Lexical-Functional Syntax. Oxford: Blackwell.
Clahsen, Harald. 1984. The acquisition of German word order: A test casefor cognitive approaches to l2 development. In R. Anderson (Ed.), Secondlanguages, p. 219-242. Rowley, MA: Newbury House.
Jordan, Geoff. 2004. Theory Construction in Second Language Acquisition.Amsterdam: John Benjamins.
Levelt, W.J.M. 1989. Speaking. From intention to articulation. Cambridge,MA: The IT Press.
Mansouri, Fethi. 2000. Grammatical Markedness and information Processing inthe Acquisition of Arabic as a Second Language. München: Lincom Europa.
Meisel, J.M., Harald Clahsen and Manfred Pienemann. 1981. On determiningdevelopmental stages in natural second languages acquisition. Studies inSecond Language Acquisition, 3, 109-135.
Pienemann, Manfred. 1984. Psychological constraints on the teachability oflanguages. Studies in Second Language Acquisition, 6 (2), 186-214.
Pienemann, Manfred and Malcolm Johnston. 1987. Factors influencing thedevelopment of language proficiency. In D. Nunan (Ed.), Applying SecondLanguage Acquisition, 21, 383-420.
Pienemann, Manfred. 1998. Language Processing and Second LanguageDevelopment: Processability theory. Amsterdam: John Benjamins.
ABOUT THE REVIEWER
Robert Felty is a graduate student in Linguistics and German at theUniversity of Michigan. His research interests include lexical access,phonetics and phonology, second language acquisition, and GermanicLinguistics.
|