• BRIDGING THE GAP BETWEEN RECEPTIVE AND PRODUCTIVE DEVELOPMENT WITH MINIMALLY VIOLABLE CONSTRAINTS

• BRIDGING THE GAP BETWEEN RECEPTIVE AND PRODUCTIVE DEVELOPMENT WITH MINIMALLY VIOLABLE CONSTRAINTS

(A REVIEW ON THE PAPER) 

• PATIOU IOANNA

• UNIVERSITY OF ESSEX, DEPARTMENT OF LANGUAGE AND LINGUISTICS, GRADUATE PROGRAMME IN LANGUAGE DISORDERS

• COLCHESTER 2004
•  

                        The child’s phonological development is composed by two domains: the perception and the production. Researches in this section have shown that there are gaps between those two fields and that the acquisition of receptive ability is always followed by the acquisition of the productive ability. This claim that “perception generally precedes production was supported by Edwards, Menyuk and Anderson and Zlatin and Koenigsknecht”. (Yeni-Komshian G.H., Kavanagh J.F., Ferguson C.A. 1980, p 119).

        But apart from their differences they should be considered as parts of the child’s linguistic competence and phonological acquisition. Furthermore, reception and production present parallel ways in development and structure, because the same markedness constraints apply in both domains. The difference and the gap between these two fields emerges in case that the representations of perception is more marked than the representations of production. This could be explained according to the process of the perceptive faithfulness constraints. Generally, the Optimality theory is the key element which will offer an integrated explanation for this gap as it views “the Universal Grammar as a set of violable constraints”. (Archangeli D. & Langendoen D.T. 1998, p 11).

•  

•   
• The development of receptive ability

          During the process of phonological acquisition, it is logical for the child to have gaps between the perception and the production. Although, some linguists have different point of view. This issue will try to present that there are indeed some gaps during the process of perceptual and productive acquisition  and some structural constraints which are similar and occur first in perception and then in production. Generally, “the child’s perception of the phonological system is complete by the time the child begins to acquire productive control of phoneme contrasts”. (Yeni-Komshian G.H., Kavanagh J.F., Ferguson C.A. 1980, p 118)   

·        Segmental constraints in early perception :                             

           According to detailed researches, Barton cites that the form of sound or meaning pairing and the different sounds that are stored in memory are components of a phonological contrast. In his experiments he tried to prove that the “ability to distinguish various contrasts emerges in a relatively fixed sequence”. (Pater J. 2004, p 221) he reaches to the conclusion that the phonemic perception – “the ability to pair a segmental phonetic contrast with a meaning difference” (Pater J. 2004, p 221) – may be developed in stages and that the ability to distinguish minimal pairs is not necessary in order to perceive a contrast.         

           Another experiment which was held by Stager and Werker was based on the concept that the infants are “capable of detecting an acoustic change in repeated stimuli”. (Yeni-Komshian G.H., Kavanagh J.F., Ferguson C.A. 1980, p 148) They showed that the pairing: sound-meaning, may result in decrease of attention to the phonemic detail. The experiment was done in infants and it was focused on the habituation/dishabituation of the sound combined with the meaning. The infants had to watch to some colored objects that were moving on a video screen while they would hear a novel syllable. This was done many times until habituation was achieved. Then, the researchers compared the mean looking time where the pair sound-meaning was diverted from the habituation phase, with the mean looking time where the pair was the same. The outcomes revealed that in case that the articulation of the initial consonant of the syllable was the only difference, the looking time was quite the same in both trials. But in case that the differences in syllables where prominent, the result was that the looking time in the switch trials was higher.         

          In addition to this experiment, another one was held with a chequer-board pattern instead of the video pattern, in order to prove that the infants didn’t perceive the consonantal place distinction in the previous experiment. That was a genuine research which was testing the perception ability of the infants and demonstrated that they can perceive a contrast. It should also be referred that the dependent variable in these experiments is the looking time.        

        In general, the overall conclusion we can arrive at, is that the consonantal place distinction is not encode in the lexical representations of the infants, although it is present in the phonemic representation. “The lexical representations of speech perception come from the same stored lexicon that is used for speech production”. (Cole R.A. 1980, p 248) Moreover, the last experiment reveals that 14 month-olds are able to perceive the phonetic level while the first experiment shoed that they can’t perceive the lexical representations.         

          In disagreement with the results of the research made by Werker and Stager, is the study of Pater et al. He used the same methodology with Werker and Stager in 14 month-olds and found that the infants didn’t notice any difference between the sub-minimal status of bi and di pair, as Werker and Stager had reported. Pater also concluded that the14 month-olds didn’t realized any voice distinction in the lexical representations.        

         Shvachkin also examined children’s ability to hear phonemic distinctions. In an extensive research with infants he reached to the conclusion that “the development phonemic perception was gradual and that it was consisted of six levels: distinction of vowels, distinction of presence of consonants, distinction of sonorants and voiced stops, distinction of palatalized and nonpalatalized consonants, distinction of sonorants and distinction of obstruents”. (Ingram D. 1976, p 23)         

            The main theme of research in this area is if the misperception is the basic reason of “some neutralizations of adult contrasts in children’s production”. (Pater J. 2004, p 224) The outcome of the experiments indicates that there are indeed lexical representations for distinctions which are lost in production and as far as it concerns the difficulty to perceive contrasts, the researchers refer that they are first being represented lexically and then being produced. Reaching to an end of this chapter, it should be cited that the lexical representations which children acquire are only segments of what they can perceive while at a later stage “when these lexical representations have been enriched, the productive representations need only to be elaborated”. (Pater J. 2004, p 225) This is in agreement with Waterson’s point of view, who proposes that “children tend to perceive utterances and reproduce the most salient features of the utterance”. (Yavas M.S. 1991, p 23)          

           Furthermore, Peizer and Olmsted and Salus and Salus support that the sounds should first be perceived in order to be pronounced correctly. And consequently, they refer that “sounds which are not pronounced later are also sounds that are perceived later”. (Ingram D. 1976, 24).

•             
• Prosodic constraints in early comprehension   
•  

         In order to explain the restrictions that occur in prosodic phonology of production it should be mentioned that some words in English are constrained by a single initially stressed disyllable, called trochaic foot which remain intact in comparison with the final stressed disyllables that lose their stressed initial syllable.

        The restrictions subjected to the trochaic foot, is being represented below for the word garbage:

      W

     F

΄σ        σ

As far as it concerns the production of the word garbage, it should be represented either with an iambic (right headed) foot:

       W               

      F          

σ        ΄σ  

or with the completion of the initial syllable to the W level:

    W

             F

 σ           ΄ σ

          Jusczyk’s researches have revealed that such constraints occur not only on children’s productive ability but also on their perceptive ability. According to his studies, where he tested 7,5 and 10 month-olds, he indicated that the first group listened far more time to passages which contained familiarized trochaic words, in comparison with the second group who listened longer to familiarized iambic words. Thus, the overall conclusion we arrive at from this experiment is that the trochaic pattern is acquired before the iambic. Extended this conclusion, we can also conjecture that there is indeed a gap between comprehension and production as it had been referred at the beginning. All these tests have clearly indicated that there are parallel restrictions occurring, in children’s perceptive and productive ability during the acquisition of phonological process.  

·        Minimal violation across perception and production                            

            One of the basic restrictions of truncation is the word size – “a word is made up of a single trochee” (Pater J. 2004, p 227) – . Another one is the violation of lexical structure – faithfulness [faithfulness constraints require the output to be identical to the input and its’ aim is to “prohibit deletion, insertions, feature changes or other changes in mapping from inputs to outputs” (Kager R., Pater J., Zonneveld W. 2004, p 20)] – called Max. More precisely Max means that “every segment of the input has an identical correspondent in the output”. (Archangeli D. & Langendoen D.T. 1998, p 63) The relationship between word size and Max constraints is the following: word size»Max: production. A rank like this would be optimal for a young learner in comparison with an adult where the rank would be reversed.         During the acquisition of receptive competence, the phonological grammar intervenes in order to regulate the structure – the complexity of the perceptive representations. That should be considered as a component of the learning process.        

             But with a strategy like this, constraints and violations restrict the lexical representations of the grammar. Thus, the grammar “intervenes between the incoming data and the stored representations, as well as taking its usual place between the stored representations and the concatenated utterance”. (Pater J. 2004, p 228) So the phonetic form of a word passes through the grammar, once in the perception and once in the production. Consequently, the input to the grammar is the perceived form and the output is the lexical form. Thus we have the same ranking in the perception, as we had in the production: word size»Max: perception.      

            Furthermore, it should be referred that the lexical representations are also subject to prosodic structure. According to the Optimality theory, the phonotactic  generalisations are the result of the interaction between the output of the markedness and the output of the faithfulness constraints.        

       Another interesting account, which is worth noticing, is that a constraint may be obeyed in a domain, but not in another. For example, the word size constraint is overcomed in perception but not in production. Prince called this phenomenon: nonuniformity, meaning that a markedness constraint may be violated only in one certain section. Nonuniformity is part of the Optimality theory. Faithfulness constraint is the responsible restriction that prevents a markedness constraint having uniformity.        

           As far as it concerns the faithfulness restrictions, they may apply either to the input-output faith or to the base/reduplicant faith. So, the structural limitations of the unmarked in reduplication applies to the following rank: faith (input-output) » markedness » faith (base/reduplicant). Taking all these into account, we reach to the conclusion that the nonuniformity’s rank is the following: faith»markedness»faith.        

           In order to be able to explain the differences in the complexity of the structures between the perception and the production, we should accept the fact that the faithfulness constraints may apply either the lexical – to the surface faith of production or the surface – to the lexical faith of perception. So there are two kinds of Max violation: the Max (LS), where “if the input is a lexical form, every segment of the input has a correspondent in the output” (Pater J. 2004, p 230) and the Max (SL) where “if the input is a surface form, every segment of the input has a correspondent in the output”. (Pater J. 2004, p 230) During the process of perceptual acquisition of the initial unstressed syllables the word size constraints exists between the Max (SL) and the Max (LS) restrictions. As far as it concerns the perception, the dominance of surface form is responsible for the representation of the initial unstressed syllable in the lexical form, in violation of word size, while in production the surface form is not dominant, resulting in an optimal truncation.        

          So, the gap between perception and production can be considered as an instance of minimal constraint violation. (Tohkura Y., Vatikiotis-Baterson E., Sagisaka Y. 1992) The role of the word size constraint changes through the levels of development as a result of the position of faithfulness constraints.       

           Thus, when the word size constraint is fully satisfied in perception and production, we get: word size»Max (SL), Max (LS), when it is minimally violated in perception we get: Max (SL)»word size» Max (LS) and when it is violated in both domains we get: Max (SL), Max (LS)»word size. The final stage is the adult level, which children have to get through this ranking until they acquire it and understand the grammar. 

·        Further articulating the model      

        Jusczyk and Aslin’s research is in disagreement with the results of Stager and Werker’s research. The first ones find out that the infants are able to represent perfectly relatively phonological detail in memory. Stager and Werker, cite that the infants link the phonological form with the meaning, to the lexical representation level, in comparison with Jusczyk and Aslin who report that the surface phonological representation is meaning-free. (Pater J. 2004) So, what they suggest is the following rank of representations and faithfulness constraints:

•  “Acoustic representation                       Present at birth  
• Faith (AS=acoustic representation
• to surface representation)    
•    
• Surface representation                            6 – 9 months   
• Faith (SL)
• Lexical representation                            11-18 months   
• Faith (LS)
• Surface representation                           18-24 months”

• (Pater J. 2004, p 233)

          In the surface representation between 6-9 months, the faithfulness constraint, faith (AS) is promoted above the markedness constraint. In the next stage of lexical representation between 11-18 months the faithfulness constraint, faith (SL) is forced to dominate the markedness constraint. While in the last stage of 18-24 months the production of words with meaning results in the dominance of faith (LS) above the markedness constraint. This is the process of acquisition that infants follow from perception to production. 

·        Comparison with other approaches: 

·        Smolensky’s model:                    

           According to this approach, the choice of the lexical forms of the candidates was made on the basis of withdraw to faithfulness. Smolensky’s proposal will show that the truncation which has already been discussed previously is optimal and so we get: word size» faith. In production the word size violation is satisfied due to faithfulness violation. In comprehension, the word size will surely be violated in the surface form. This means that the markedness constraints – which are “the natural antagonists of faithfulness constraints because faithfulness may preserve the lexical properties while the markrdness may ban them at the surface” (Kager R., Pater J., Zonneveld W., 2004, p 20) – will create an iambic foot without taking into account the initial syllable.       

           Smolensky’s approach reveals that the comprehensive ability is advanced relative to the productive ability. But the lexical forms of comprehension should be perfectly phonological comprehensive. This is impossible to occur and it could be explained in terms of the infant’s comprehensive and productive ability which are subject to similar constraints.   

 ·        A mixed model  :              

             As it has been referred in a previous section, the comprehensive process of acquisition is influenced by the markedness constraints, since the lexical form is considered to be the output. But this proposal rejects the equivalence either between the lexical form and the input or the surface form and the output. The following approach of mixed model will try to explain and restore this equivalence. The basic element in this model is to accept that the acoustic-phonetic representation produces the phonological surface string. This could be done only by considering the acoustic-phonetic string as output to the grammar.

         Apart from this, the mixed model should also be combined with Smolensky’s view that “the surface strings must faithfully represent the phonetic form”. (Pater J. 2004, p 236) The three stages of development during the comprehensive acquisition are similar with the stages of Minimal violation across perception and production section. The only difference is that instead of faith (AS), here we will use faith (SL). So the process of acquisition is subject to the following stages:

• Stage 1: markedness » faith (AS), faith (LS)
• Stage 2: faith (AS) » markedness » faith (LS)

• Stage 3: faith (AS), faith (LS) » markedness

              The first stage represents reduced representations in both domains of comprehension and production, the second stage reveals faithful comprehension and reduced production and the last one shows faithful comprehension and production.         

           It should also be referred that during the acquisition of comprehension and production the words are restricted to a single trochaic foot. This limitation occurs because “the structural constraints apply to the surface phonological form constructed on the basis of the perceived acoustic-phonetic representation”. (Pater J.2004, p 236)         

             The significant point of this model is that it allows the: faith (AS) » word size » faith (LS) ranking happen, producing by this way, the gap between the perceptive and the productive ability. As far as it concerns the comprehension, the faith (AS) is satisfied by the markedness prosodic structure, while in production we have the unmarked structure as a result of the continuous supremacy of word size over faith (LS). Moreover, the general conclusion we arrive at from this model is that the restrictions in the structure, influence the perceptive outcome.          Furthermore, in order the input –output mapping to be equivalent with the lexical–surface representations, we should be based on the Optimality theory.  

·        The dual-lexicon model  :                     

       The general concepts of the dual-lexicon or two-lexicon model are the separate representations for production and perception and the different faithfulness constraints that occur in every domain. The markedness constraints result in difference in the ranking of faithfulness limitations and in restriction to the extention of diverge. Although, another more significant difference emerges: when a knowledge has been acquired from the receptive ranking of markedness constraints, it is transferred immediately to the production grammar and so it doesn’t need to be relearned.         Bringing this chapter to a close, we should refer that the Optimality theory gives the answer to the process of child’s output representations: the output limitations interact with the lexicon with the help of the faithfulness constraints and this leads the child to the creation of his output representations. For example the word: fish: “its’ perceived form is entered first in the input lexicon. Then the reduction rules, constraints, of child’s phonology, turns this form into one underlying production. I and S are stored in the output lexicon as an unordered pair of vowel and fricative which must be ordered by the production rules”. (Kager R., Pater J., Zonneveld W. 2004, p 13) Then it is subjected to the output constraints that fricatives are not allowed to be positioned in onsets. So when a new phonological rule is being acquired by a child, the existing pronunciations may persist, “as if output forms serve as independent lexical items”. (Menn and Matthei 1992, p 213 in Kager R., Pater J., Zonneveld W.2004, p 14)  

·        Conclusion:                      

            Smolensky proposed to use Optimality theory in order children to acquire the receptive and productive phonology with a single grammar. The differences that emerge are the constraints in the early comprehension and production, where the receptive phonology is acquired perfectly from the beginning.          

          This issue also revealed that the receptive ability is being developing in a similar, parallel way to the development of productive ability.         

            Although, a series of structural limitations ranked over faithfulness constraints, apply to both domains production/reception and surface form/lexical form. So children acquire the reception by “the promotion of comprehension specific faithfulness constraints” (Pater J.2004, p 239) while they get their productive competence later, by the acquisition of lower rank of the production specific faithfulness constraints.         

           The overall conclusion we can arrive at is that the markedness constraints in perception, intervene between the listening signal and the lexical representation. In addition to this, researches have shown that the phonotactic limitations play also an important role in adult’s speech perception.        

            Bringing this essay to a close, we should cite that it was a worthy attempt to present in a simple and well-structured way the process of phonological acquisition and offer a detailed explanation for the gap between perception/production. Generally, this issue is an integrated paper which presented new researches and shed new light upon the subject of bridging the gap between reception/production.