关键概要critical summary

As for grammar parts, persons’ good appearance demonstrate that no matter quantity of inconsistent determiners , there is little influence for their learning. Next to sentence tests, we can easily find that with the help of frequent molds, persons from noisy circumstances are more likely to regularize the input than that from control ones. Then for judgments parts, sentences which have wrong positions for determiner are hated by all, and those participants never  heard are rated lower than the less usually ones. In general, a prevailing phenomenon is that with more and more noisy determiners arising, participants will be more likely to enjoy the main determiners parts than the less usually ones.

 

总而言之，试验清楚地表明，正规化，可以指导和成年人实现。然而，显著结果仅发生在那里的情况很多不安嘈杂的限定词，它激励着我们要强调另外两个因素显著。All in all, the experiment clearly indicated that regularization could be guided and achieved by adults. However ,  significant results only occurred in the circumstance where many noisy determiners disturbed, it inspired us to emphasize another two significant factors.
 

首先，显著正规化可以是较不频繁的嘈杂限定词不为特色的自然语言现象不一致的人的结果。要解决这个问题，我们设计了实验2.其次，同样的现象是否也会发生儿童。更准确的说，什么孩子的程度和趋势，以规范不同于成人的投入。所有上述问题将在实验3进行调查。Firstly, the significant regularization may be the result of less frequent noisy determiners not for  inconsistent ones featured natural language phenomenon. To resolve this ,we designed experiment 2. Secondly , whether the same phenomenon will also happened for children. More accurately, what the extent and trend for children to regularize the input different from adults. All the questions above will be investigated in experiment 3.    

Body:主体

In order to get the explanation and conclusion about this kind of regularization, three experiments were carried out. Some clearly, defined targets are put forward before experiments, including how learners learn languages under disturbance of inconsistency or probabilistic trend, at what time and for what reasons learners get the abnormal language as regular and natural as their learned ones, in the process of reproducing, adults are tendency to realize the complexity and characters of inconsistence before accomplishing regularizations, contrary, children are not.

In the followings, I will take a specific glance on experiment 1, to summarize the methodology and main findings in detail and try to introduce the findings as clearly as possible.

In experiment 1, the focus is that if input filled with complexity is regularized more than the simple ones also described as absence input by adults. For the sake of investigation, an artificial language paradigm is applied, and then participants are exposed to abnormal input with inconsistency cannot be predicted, provided with much more alternatives, this kind of input is identified as more complex than former ones. 
 

The methodology is designed aimed for five different parts: participants, the language, presentation, experimental manipulation, tests. In the first place, participants are all adults and attendants from the same university, of course, they are paied for their devotion to the experiment. Secondly, what presented before the participants was languages formed by small basic words, combined with videotapes describing things and performence. Under the premise that, the sentences are able to meet the basic application for the realistic world, although  the combination will be limited to the group of small words. Thirdly, there are 9 videotapes (including 1 for testing) provides languages that showed for participants. They are required to report again the sentence which are showed on the videotape. In order to ensure they accomplish the sentences on their own, they are informed it is only a test for pronunciation. Subsequently, there are many measures applied to make sure the equality of the experiment, such as match quantity of subjects, the same conditions designed for persons. Under these circumstances, the only difference is the appearance of determiners. Besides, there are 4 various tests for evaluation their actions. 2 vocabulary tests were given but few persons realized the second performance and record are equivalent to the first. One of the most interesting portions is completing the sentence. By the way, we could estimate the inconsistent portions like determiners. There is another assistant test to evaluate what a person knows about the use of determiners, called judgment task. Be careful to notice the difference between judgment and production. When it comes to the remains of the input, there is also a judgment test for the sake of general grammar.
 

After the scientific methods, we reached the primary results. According to vocabulary results, it is a little disappointed that we do not find difference significantly. As for grammar parts, persons’ good appearance demonstrate that no matter quantity of inconsistent determiners , there is little influence for their learning. Next to sentence tests, we can easily find that with the help of frequent molds, persons from noisy circumstances are more likely to regularize the input than that from control ones. Then for judgments parts, sentences which have wrong positions for determiner are hated by all, and those participants never  heard are rated lower than the less usually ones. In general, a prevailing phenomenon is that with more and more noisy determiners arising, participants will be more likely to enjoy the main determiners parts than the less usually ones.

Discussion:讨论

All in all, the experiment clearly indicated that regularization could be guided and achieved by adults. However,significant results only occurred in the circumstance where many noisy determiners disturbed, it inspired us to emphasize another two significant factors.

Firstly, the significant regularization may be the result of less frequent noisy determiners not for  inconsistent ones featured natural language phenomenon. To resolve this ,we designed experiment 2. Secondly , whether the same phenomenon will also happened for children. More accurately, what the extent and trend for children to regularize the input different from adults. All the questions above will be investigated in experiment 3.

Despite all the scientific and systematic experiment methods , there is one thing in experiment 1 that is sort of weak. From my perspective, the foundation theory of the experiment is Universal Grammar. There are still remain lots of controversy for the theory ,whether human ability of language is born,or acquired; the evolution of language is the result of human genes, or historical and cultural influence. Recursion plays an crucial role in acquiring language, if there is no similar expression in a nation's existing language system, then regularization will absolutely not happen .These factors were missed when choosing the participants, the balance of gender, the background of culture and other things varies in nations. Therefore, the controversial and debatable theory of Universal Grammar lead to lack of consideration and rigorous application.

阅读的原文：When natural language input contains grammatical forms that areused probabilistically and inconsistently, learners will sometimesreproduce the inconsistencies; but sometimes they will insteadregularize the use of these forms, introducing consistency in thelanguage that was not present in the input. In this paper we askwhat produces such regularization. We conducted three artificiallanguage experiments, varying the use of determiners in the typesof inconsistency with which they are used, and also comparingadult and child learners. In Experiment 1 we presented adult learnerswith scattered inconsistency – the use of multiple determinersvarying in frequency in the same context – and found that adultswill reproduce these inconsistencies at low levels of scatter, butat very high levels of scatter will regularize the determiner system,producing the most frequent determiner form almost all the time.In Experiment 2 we showed that this is not merely the result of frequency:when determiners are used with low frequencies but inconsistent contexts, adults will learn all of the determiners veridically.In Experiment 3 we compared adult and child learners, findingthat children will almost always regularize inconsistent forms,whereas adult learners will only regularize the most complexinconsistencies. Taken together, these results suggest that regularizationprocesses in natural language learning, such as those seenin the acquisition of language from non-native speakers or in theformation of young languages, may depend crucially on the natureof language learning by young children.[1]2009 Elsevier Inc. All rights reserved.0010-0285/$ - see front matter [1]2009 Elsevier Inc. All rights reserved.doi:10.1016/j.cogpsych.2009.01.001* Corresponding author.E-mail address: clhudson@berkeley.edu (C.L. Hudson Kam).Cognitive Psychology xxx (2009) xxx–xxxContents lists available at ScienceDirectCognitive Psychologyjournal homepage: www.elsevier.com/locate/cogpsychARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.0011. IntroductionIn the past half-century, great strides have been made in documenting the linguistic accomplishmentsof children as they learn their native languages. Despite this increased base of knowledgeregarding what learners know as their language abilities develop, we are only beginning to understandhow this learning takes place. This paper addresses questions about the learning mechanisms themselves,focusing on the limits of the system by examining cases where learners acquire something differentthan the patterns in their input.We examine what human language learners can (and cannot) acquire when the input is abnormalin a particular way. Occasionally, language learners are exposed to input that contains grammaticalinconsistency, what we call probabilistic grammatical tendencies: a form is used some percentage ofthe time in a particular context, with its occurrence not predictable on the basis of any features ofthe context. This kind of input is unusual in that probabilistic grammatical tendencies of this sortare not typically found in human languages. However, they do occasionally occur, for instance whenlearners are acquiring their language from non-native speakers. Evidence suggests that this unpredictablevariation disappears as the language is learned; it is regularized (Newport, 1999; Ross & Newport,in prep; Singleton & Newport, 2004). The result of this change is a language that is no longer abnormal;the language as spoken by the learner is like other natural human languages. Learners, it seems,are able to ‘fix’ or repair this kind of abnormal input.Here we examine how this change is accomplished, by examining the kinds of information that humanlearners extract from inconsistent linguistic input. Our interests are broader, however. We are notsimply interested in characterizing the learning mechanisms involved in abnormal situations; we areinterested in gaining a greater understanding of the mechanisms involved in language acquisitionmore generally. We submit that understanding the performance of learning mechanisms when the inputis atypical can contribute to our understanding of the learning mechanisms involved in typicallanguage acquisition as well; the input may be unusual, but the learners are not.We present results from three experiments investigating how humans learn from languages containinginconsistent or probabilistic grammatical tendencies, asking about the circumstances underwhich they succeed at learning the variation veridically, and when and why they make the languagemore regular and more like other natural languages as it is learned. In previous research we haveshown that, at least in one circumstance, adults reproduce the inconsistencies they are exposedto, while children do not. In the present work we investigate this question more comprehensively,examining both the age of learners and also the nature and complexity of the inconsistencies towhich they are exposed, to see how these variables affect what is learned. In Experiment 1 we presentadult learners with inconsistency that is either relatively simple or more complex, to investigatewhether they might be less likely to learn veridically, and more likely to regularize, when the inconsistenciesare complex. In Experiment 2 we present adult learners with a complex but consistentlanguage, to see whether complexity in and of itself is enough to induce regularization in adultlearners. In Experiment 3 we vary the age of the learner, testing both adults and children to investigatewhether children are more likely to regularize a range of types of inconsistencies than areadults. To foreshadow, our results suggest that humans can learn from inconsistent linguistic input,but also that they do indeed make it more consistent under certain circumstances. Importantly, thedegree to which they regularize depends on both the age of the learner and the presence and natureof the inconsistency. In the discussion we will return to the question of how our findings fit intobroader issues, and particularly what we think our findings say about the mechanisms involvedin language learning and the circumstances under which natural languages become less inconsistentand more regular.1.1. Linguistic variation and accompanying change by learnersTypically, language learners are exposed to input that contains very consistent grammatical patterns.Sometimes these patterns are deterministic: a particular grammatical form is used every timea particular meaning is expressed. The regular plural form in English is an example of a deterministic2 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001pattern (though even such patterns can have lexical exceptions).1 Nouns take –s, -z, or –iz in the plural,depending on the phonological form of the noun (and those that are exceptions are always exceptional,in every context in which they occur). Other patterns are variable but are nonetheless grammatically predictable:the same form does not always occur with a particular meaning, but the variation among formsis contextually dependent and consistent across speakers (Labov, 1969, and papers in Chambers, Trudgill,& Schilling-Estes, 2003). As Smith, Durham, and Fortune (2007) put it, the alternation between forms is‘variable but highly structured’. For example, the pronunciation of –ing in English varies between ‘–ing’and ‘–in,’ with ‘–in’ being the casual form, used in faster speech, less formal contexts, and more often byyounger rather than older speakers. However, sometimes learners are exposed to linguistic input thatcontains grammatical patterns that are truly inconsistent tendencies. These patterns are unpredictableand probabilistic in nature; a form is used some percentage of the time in a particular context, withits occurrence or non-occurrence not predictable on the basis of any features of the context (see e.g.Newport, 1999; Singleton & Newport, 2004). We call these inconsistent or probabilistic grammaticaltendencies.Probabilistic grammatical tendencies are very common in the speech of late learners of a secondlanguage (Adamson, 1988; Goldowsky & Newport, 1993; Johnson, Shenkman, Newport, & Medin,1996; Newport, 1990, 1999). Late learners are generally not as proficient as early learners. The mostnoticeable difference is often their accent, but late learners also have problems with grammatical deviceslike tense and aspect, agreement marking, and case marking (Birdsong, 1999; Johnson & Newport,1989). Speakers who have learned a language as adults may simply omit the grammaticalmarking altogether, but often they will use a grammatical device inconsistently and probabilistically.This probabilistic usage can take different forms: the probabilistic usage of a single form; an unpredictablealternation between several forms (only one of which would be considered correct in the nativeform of the language); or a combination of the two, with the speaker sometimes using the correctform, sometimes an incorrect form, and sometimes using no form at all. Probabilistic usage is seen insecond language interlanguage (during learning, Kanno, 1998), as well as in fossilized asymptotic secondlanguage grammars (Sorace, 2000). Thus, although probabilistic usage is atypical of mature nativespeakers, it is not simply a characteristic of language learning in progress. Furthermore, the specificprobabilities and patterns of usage differ between individual second language speakers, even whenthey share the same native language, with the result that there may be no consistency across speakerswithin the same community (see, for example, Meisel, Clahsen, & Pienemann, 1981; Wolfram, 1985).A question of interest is therefore how children might behave if they had to acquire their own nativelanguage from parents (or other adults) whose usage contained such inconsistencies. Children areknown to have difficulty acquiring lexical exceptions to grammatical rules (Marcus et al., 1992). Linguisticinput of the kind described above, with true inconsistencies, seems as if it should be particularlydifficult to learn from. There are a few recorded instances of learners facing just this kind of input(see e.g. Aitchison, 1996; Kotsinas, 1988; Newport, 1999; Sankoff, 1994; Sankoff & Laberge, 1973; Singleton& Newport, 2004). The outcomes suggest that, while learners can acquire language from thiskind of input, they do not acquire such inconsistent variation veridically. In contrast with the moretypical situation of structured variation in the input, which is learned correctly (Kovac, 1981; Labov,1989; Roberts, 1997; Smith et al., 2007), learners exposed to inconsistent input appear to change thelanguage as they learn it, making it more regular.1.1.1. Children learning from non-native inputSingleton and Newport have conducted very detailed studies of the acquisition of American SignLanguage (ASL) by a deaf child they called Simon, whose only input source was his deaf parentswho were late learners of the language (Newport, 1999; Ross & Newport, 1996; Singleton, 1989; Singleton& Newport, 2004). Like other late learners of ASL, the parents’ signing contained many errorsand was governed by probabilistic, rather than deterministic, rules. That is, they would use complexmorphemes each some percentage of the time in the obligatory context, with its occurrence or non-1 Although there is variation in the pronunciation of the plural form (cats versus dogs versus canvases), it is completely consistentand predictable, and depends on phonological form of the noun.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 3ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001occurrence inconsistent and not predictable on the basis of features of the context. Simon did notreproduce the unpredictable inconsistency present in his input, however; he changed the system,using the forms most frequently present in his input almost categorically. This is most compellinglydemonstrated by one particular regularization that Simon made. His parents most frequently useda handshape for vehicles that is not typical in ASL; Simon used their incorrect handshape consistently.As Singleton and Newport suggested, this result indicates that he is regularizing his parents’ inconsistentsystem, and not somehow secretly learning ASL from another source (Singleton & Newport, 2004).He received variable input and made it far more consistent and predictable. Ross and Newport (1996and in preparation; Ross, 2001) verified this claim in longitudinal analyses of Simon’s ASL usage, andalso demonstrated a similar outcome in three other deaf children learning ASL from their late-learninghearing parents.Kotsinas (1988) reports on a similar case involving children who were immigrants to Sweden.These children lived in immigrant communities in Stokholm and learned their Swedish primarily fromtheir parents and other immigrants to the community, all of whom were late learners of Swedish.According to Kotsinas, the parents’ speech contained ‘‘considerable variation among the speakers’varieties” (p. 133). The children’s productions, however, displayed more consistency, indicating theemergent grammaticalization of some of the forms present probabilistically in the parent’s speech.Importantly, the non-standard Swedish the children spoke was not the same as the vernacular Swedishthat ethnic Swedes speak – it was a modified version of the pidgin-like variety spoken by their parents.Notably, although some of the children also spoke Standard Swedish learned through exposure atschool, their vernacular was well stabilized prior to exposure to the standard variety.1.1.2. The language of the first generation of native speakers of a new languageSimilar phenomena also occur during the emergence of a new language. A number of researchershave documented the changes that occurred in two related pidgin languages as they were acquired bythe first generation of native speakers. The first, Tok Pisin, a contact language spoken in Papua NewGuinea, has been extensively studied by Sankoff and her colleagues, among others. They were particularlyinterested in the presence and form of grammatical devices (such as tense and aspect marking)and the clause and sentence structure in the speech of those who learned the language as adults, ascompared with the speech of those who learned the language as a native language.Though the language was in the early formational stages, they did find grammatical structures inthe speech of the adults; but they also found some variability in the use of those structures. This variabilitytook several forms, with two of these most relevant to the present work. First, as would be expectedin a late-learned language, there was variation in occurrence: any particular form occurred inits appropriate context probabilistically (Aitchison, 1996; Sankoff, 1994).2 Second, there were meaningsthat could be expressed in any of several ways (Sankoff, 1979).3Importantly, the speech of the children learning Tok Pisin as their native language from these latelearningmodels contained less unpredictability (Romaine & Wright, 1987; Sankoff, 1979, 1994; Sankoff& Laberge, 1973). For instance, although the native speakers produced the preverbal form i in thesame locations as did their non-native-speaker elders, the frequencies of use in the various syntacticenvironments differed between the children and adults (Aitchison, 1996; Sankoff, 1994).Similar results have been found in Solomon Islands Pijin (Jourdan & Keesing, 1997), anotherdescendant of Melanesian Pidgin (Keesing, 1988), and these kinds of changes have also been proposed,although not directly witnessed, in French-based creoles (Becker & Veenstra, 2003).2 It should be noted that there was also variation between speakers that was semi-predictable (Sankoff, 1994). Some of theinterspeaker variation was conditional on age, such that speakers who learned the language at different times spoke a littledifferently from each other. Some of it was correlated with location of origin, such that speakers from different regions of thecountry had different typical speech patterns. There were also sex differences. However, even beyond this, there was a great deal ofunpredictability in the speech of any one individual.3 Non-creole languages of course also often possess multiple ways of expressing the same meaning. For instance, the Englishfuture tense can be expressed using ‘ to be going to’ or ‘will’, or in short colloquial forms ‘be gonna’ or ‘’ll’ (e.g., I am going to go toSouth Carolina, I will go to South Carolina, I’m gonna go to South Carolina, I’ll go to South Carolina). These are not truly in freevariation with each other, however; they differ in certainty and formality, and possibly in focus. The Tok Pisin forms are not in freevariation with each other either, but they differ more according to context. See Sankoff (1979) for a more complete discussion.4 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.0011.1.3. Regularization by adult learnersThus far, all the examples we have given involve regularization by children. However, there are reasonsto believe that adults may also regularize variability in languages. Aitchison (1996; Aitchison &Agnihotri, 1985) points out the tendency of adult language learners to overregularize morphological aswell as syntactic patterns. For example, one adult learner of German described by Klein and Perdue(1993) always used eine as the indefinite article (when he used an article with indefinites), despitethe fact that German has multiple indefinite article forms that vary by gender of the noun (as doesItalian, his native language). This can be seen as a reduction or regularization of variation (even thoughthe variation in German is predictable). Unlike overregularization in children’s first language acquisitionfrom native input (e.g. the overregularization of –ed to irregular verbs), much of the adult learner’soverregularization remains in their system as it fossilizes; they do not outgrow it (Adamson,1988; Sorace, 2000).1.1.4. What kind of mechanism is likely to produce regularization?As presented, the evidence suggests that both adults and children can, at least in principle, introducegreater regularity into languages. Are the learning mechanisms that produce regularization specificto language learning, or might they be more general in scope? Some results from studies ofprobability learning suggest that it might be the latter. The general aim of probability learning studieswas to describe what participants learn when they are provided with information that is probabilisticin nature. For instance, participants are asked to watch two lights that flash, one at a time. The participant’stask is to make a prediction about which of the two lights will flash just before each flash event.Which light actually flashes is probabilistically determined so that the overall probability is within apre-determined range. For instance, in a 70/30 experiment, light A flashes 70% of the time, and light Bflashes 30% of the time, entirely probabilistically. (The particular ratio can, of course, differ byexperiment.)Most experiments in this literature show that after very little exposure, adults’ predictions begin tomatch the exposure probabilities. For example, in the 70/30 example, participants predict that light Awill flash next on 70% of the trials and that light B will flash next on 30% of the trials (Estes, 1964,1976). This kind of response pattern is called probability matching. (Note that probability matchingis not the optimal response for success in prediction or in securing reinforcement, as the paradigmis run in animals, since predicting A on 70% of the trials, when it does in fact flash unpredictably witha .70 probability, will lead the participant to be correct only 58% of the time. Nonetheless, probabilitymatching is the most common response pattern seen in these experiments.) However, under certaincircumstances another type of response appears. For instance, when participants are asked to attendto their level of correctness on blocks of trials rather than for each individual trial, they tend to overmatch,selecting the more frequent alternative more often than it actually occurs. Of particular interestto us, some experiments suggest that one can induce overmatching by changing features of the presentation.Gardner (1957) and Weir (1972), for example, found that adults overmatched when presentedwith more than two alternatives. For example, if light A flashes 70% of the time, and lights Band C each flash 15% of the time, participants guess light A more than 70% of the time. This literaturethus suggests that adults will regularize non-deterministic information under some conditions.Similar experiments conducted with young children show that they are less likely than adults toprobability match and more likely to regularize. (This is our term, which we use to suggest similarityto regularization in language acquisition. In the terminology of this literature, children are more likelyto overmatch or even to pick the more frequent item all the time, called maximizing.) However, thedegree to which they regularize, and the age at which they stop regularizing and begin to probabilitymatch, varies across studies. The general trend in the child literature is that younger children are morelikely than older children to overmatch (see discussion in Hudson, 2002). The upper limit of regularizingbehavior is dependent on the task, with more complex tasks producing regularizing at older agesthan easier tasks. Bever (1982) found high degrees of overmatching in 2- and 3-year old children in atwo-choice task, and very little maximizing in 4-year olds in the same experiment. Kessen and Kessen(1961) found probability matching by age 3;7. However, overmatching has been found in children asold as 5 and 7 using a slightly different task that included three choices (Stevenson & Weir, 1959; Stevenson& Zigler, 1958; Weir, 1964). The findings, then, are similar to those in the adult studies, whereC.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 5ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001increased complexity produces less probability matching and increased regularizing behavior. However,the literature also shows quite a bit of variation in the tasks (especially with respect to theamount of information available to the learner) and in the details of their outcomes.1.1.5. Predictions for language learningTaken together, the findings from studies of language acquisition and probability learning reviewedabove provide some suggestions about variables that might lead to the regularization that occurs inthe acquisition of non-native language input or in the emergence of a new language. These findingssuggest that the nature of the inconsistencies themselves may play a role in regularization, and alsothat children may be more likely than adults to regularize inconsistencies; but neither of these variableshas previously been investigated systematically. Indeed, as discussed above, typical human languagesdo not contain much unpredictable inconsistency, so it is difficult to answer this question byexamining normal language acquisition. There are, as discussed above, natural situations where linguisticinput is provided by non-native speakers. However, in such cases many variables are confounded,making it difficult to determine which of them lead learners to change the language. Wehave therefore developed a miniature artificial language paradigm for use in investigating this question.In previous work using this paradigm, we presented learners with simple linguistic input thatcontained one inconsistent part of the grammar. We found that adult learners did tend to reproduce(or probability match) the inconsistencies in their input; but, in contrast, children turned inconsistentforms into rules (Hudson Kam & Newport, 2005).In that study we exposed learners to an artificial language with probabilistically occurring determiners(articles, like ‘the’ and ‘a’ in English). Nouns were accompanied by determiners some percentageof the time; the rest of the time determiners were absent, and the nouns were alone in the nounphrase. When nouns had determiners and when they did not was determined randomly: there wereno differences in meaning or other aspects of context when the nouns appeared with versus withoutdeterminers. In this way the variation was unlike that typically present in natural human languages,though much like that in the speech of late learners, adult speakers of emerging contact languages,and the parents of children like Simon. At testing, adult participants produced about as many determinersin their own productions as they had heard in the exposure (that is, they probability-matchedthe use of determiners). In contrast, children’s productions were more systematic than their input.However, the type of inconsistency investigated in that study – variation in the presence versus absenceof a form – is not the only type of inconsistency that occurs in the natural language phenomenawe are interested in understanding. In the present series of studies, we examine a different type ofinconsistency, more like that found in the natural input received by Simon and by those exposed toan emerging language, to see how this type of inconsistency affects adult learning; and we also observechild learners exposed to the same type of inconsistency, to see how age differences interactwith this. In Experiment 1 we ask whether adult language learners will regularize more when the languagethey are exposed to contains more complex variation than when it contains simple two-alternativevariation. In Experiment 2 we consider in greater detail the character of the complexity andinconsistency required to produce regularization. In Experiment 3 we compare the performance ofadult learners with that of children exposed to these types of inconsistencies, to see if they regularizein the same or different ways.2. Experiment 1In this experiment we examine whether adult learners regularize complex variable input morethan the simple presence/absence input. To investigate this we used an artificial language paradigmto expose participants to miniature languages containing unpredictable inconsistency that was morecomplex than the inconsistency we had previously found to produce probability matching in adultlearners (Hudson Kam & Newport, 2005). We did this by increasing the number of options in a particulargrammatical category. (Exactly what this means is described below.) This is much like increasingthe number of lights over which participants make predictions, a manipulation that produces overmatchingin probability learning studies, and thus should produce overmatching in our language-6 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001learning task if the two result from the same mechanisms. It is also very much like the kind of inconsistencyto which Simon was exposed and which he regularized. We exposed participants in differentconditions to increasing numbers of options, reasoning that increased options equals increased complexityin the nature of the probabilistic variation and therefore perhaps increased regularization.As in our earlier studies, we exposed participants to a language in which all the elements displayedregular properties except the determiners. Determiners were selected because we wanted to study thelearning of an inconsistent functional category, and in a short period of time. This makes most otherfunctional categories unsuitable, given the complexities of the meanings they encode. As in HudsonKam and Newport (2005), we examined what participants had learned about the language using severaldifferent measures, including production and grammaticality judgment tasks.2.1. Method2.1.1. ParticipantsParticipants were students at the University of Rochester at the time of the study. Average age was19.8 years. Forty-one women and 19 men participated. They were paid daily for their participation andreceived a bonus after the final session for completing the entire experiment. They were recruited primarilyfrom the department’s subject pool list via an email describing the study and inviting them toparticipate.2.1.2. The LanguageThe basic language to which we exposed participants was small, consisting of 51 words: 36 nouns,7 intransitive verbs, 5 transitive verbs, 1 negative, and 2 main determiners, 1 for each of 2 nounclasses. These words were in the input of participants in all conditions. In addition, there were another16 noise determiners used in the experimental manipulation. These did not vary by noun class, but theexact number of noise determiners varied across the experimental conditions. (This is described ingreater detail below.)The language was presented in conjunction with a small world of videotaped events showing objectsand actions, whose permissible combinations restricted the number of possible sentences. Evenwith these semantic restrictions there were over 13,200 possible sentences in the language. Thegrammatical structure of the language is shown in Fig. 1. The basic word order is (NEG) V-S-O. Asis typical for a real VSO language (Greenberg, 1963), the determiner follows the noun within theNP. This word order was selected to be quite different from that of English, and also to permitthe use of a sentence completion task (see below) that would readily elicit NPs (the crucial portionof the artificial language sentences) from participants during testing. This basic grammatical structurepermits four possible sentence types: intransitive, transitive, negative intransitive, and negativetransitive.Although the vocabulary is relatively small in comparison to full natural languages, we took greatcare to make the language as realistic as possible. (See MacWhinney, 1983, for a discussion of thispoint with respect to research using miniature artificial languages). Complete sentences can be producedin the language, and there are different kinds of sentences (e.g. negative and positive, transitiveand intransitive). The word order and functional category properties were modeled after those of naturallanguages (though unlike English, in order to avoid simple transfer), and the sentences expressedmeanings. In addition, while some of the natural language cases we were modeling did not containtheir inconsistencies within the determiner system, the details of their probabilistic variations arevery similar to the inconsistencies of our artificial languages.Fig. 1. Grammatical structure of the language.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 7ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001The nouns were divided into two classes. Nouns were assigned to classes on a completely arbitrarybasis, with 20 nouns in class 1 and the remaining 16 nouns in class 2. This was done to keep the methodsas similar as possible to those in earlier studies to permit comparisons of previous and current results.The only grammatical consequence of noun class membership in the language is determinerselection: each class of nouns takes a different main determiner. (A word list and gloss for each wordcan be found in Hudson Kam & Newport, 2005.) This division into two noun classes, with a differentmain determiner for each, is similar to the division of nouns into two gender or declension classes inmany natural languages. The exact nature of the linguistic input received by a participant variedaccording to consistency condition assignment and is described below.2.1.3. PresentationParticipants were exposed to the language by videotape for eight sessions, each lasting 25–29 min.Participants were seated in front of a video monitor, on which they watched a scene or event. Theythen heard a sentence in the miniature language that described the scene. Sentences were spokenat a normal rate with English prosody and phonology and sounded very natural and fluent. Therewas no explicit instruction in grammar or vocabulary, and they never saw anything written. Participantslearned the language solely from the auditory exposure to the sentences. For example, a participantsaw a toy boat hitting a girl-figurine and heard:(1) /flIm m[1]wzner kA ferlukV po/hit boat DET1 girl DET2‘The boat hits the girl.’The exposure set contained 230 sentences and their corresponding videotaped scenes. Half the exposureset sentences were intransitive and the other half were transitive. Negative sentences were includedto help the participants learn the meaning of the verbs, especially the intransitives, as wellas to expand the number of possible sentences in the language. Overall, however, there were relativelyfew negative sentences in the presentation set (seven transitive sentences and 43 intransitivesentences).Verb frequencies varied due to the importance of keeping noun occurrences balanced, along withthe constraints arising from the meanings and associated selectional restrictions of the verbs. Eachintransitive verb occurred 15–18 times in the 115 intransitive sentences. Each transitive verb occurred14–27 times in the 115 transitive sentences. Individual verbs were presented either in both negativeand positive sentences or in only positive sentences; no verb was presented in only negative sentences.Each noun in the language occurred 3–4 times in the intransitive sentences, and 3–4 timesin each syntactic position (subject and object) in the transitive sentences. Like the verbs, each nouncould appear in both positive and negative sentences, or only in positive sentences; no noun appearedin only negative sentences. (The exact number of times any particular word occurred in the exposureset is listed in Appendix B of Hudson Kam & Newport, 2005).Each exposure session contained a different set of approximately 115 sentences drawn from the230 sentence exposure set. Each sentence (and scene) was presented four times over the course ofthe eight exposure sessions. Participants were asked to repeat each sentence after hearing it. Theywere told that this was pronunciation practice which would be helpful since they would have to producetheir own sentences at the end of the experiment. The entire experiment took nine sessions tocomplete (the eight exposure sessions and one test session). Participants completed the experimentin 9–12 days. All exposure and testing was done individually.2.1.4. Experimental manipulationIn this experiment as well as those that follow, equal numbers of subjects were assigned to each ofthe five conditions (input groups) described below. Participants in all five conditions were exposed tothe same basic sentences, and therefore their exposure was equivalent in almost all aspects of sentencestructure. Input sentences differed between conditions only in the occurrence of determiners8 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001in the noun phrases. All participants were exposed to sentences containing the main determiner forms(those agreeing with the class of the noun) 60% of the time. Although the sentences containing themain determiner forms were selected randomly from the exposure set, they were the same for all participants,regardless of complexity condition.The experimental manipulation occurred in the remaining 40% of noun phrases. Participants in thecontrol condition heard no determiner form in those noun phrases. These participants were exposedto the same input as in our previous study (Hudson Kam & Newport, 2005), a kind of inconsistency wecall presence/absence inconsistency. This inconsistency exhibited the least complex variation, and onthe basis of previous results participants were expected to probability match the occurrence of thedeterminers in their own usage. All other participants received more complex determiner variation,of a type we call scattered inconsistency. In scattered inconsistency there is one main form that occursa majority of the time, but a number of other forms may occur instead of this main form, inconsistentlyand at a lower frequency. The degree of scatter differed across conditions. Like the control participants,those in the 2 ND group (ND = noise determiner) heard the main determiner forms 60% of thetime, but in the remaining 40% of noun occurrences, one of two other determiner forms (hereaftercalled ‘noise’ forms) occurred, each in 20% of the noun occurrences of each noun class. For noun class1, 60% of the noun phrases occurred with the main determiner form, /kA/, 20% occurred with the noisedeterminer form /te/, and the remaining 20% with the noise form /meg/. The same was true for nounclass 2: 60% of the noun phrases had the main form /po/, 20% had /te/, and 20% had /meg/. Note thatthe noise forms occurred with both noun classes, unlike the main forms, which were restricted in theirdistribution to the main noun class. Noise forms were thus both lower in frequency and more unpredictablein the context in which they occurred.The input for the three other complexity groups was similar, but contained more noise determinersthat each occurred with lower frequency. The 4 ND group heard the main determiner form 60% of thetime and 4 noise determiner forms, /te/, /meg/, /li/, and /kum/, that each occurred with 10% of thenoun phrases within a noun class. Again, the noise forms occurred with nouns in both classes. The8 ND condition heard the main determiner 60% of the time and 8 noise forms that each occurred 5%of the time. In addition to the 4 noise forms listed above, they heard /su/, /gI/, /ler/, and /bAn/. The finalcondition was the 16 ND condition. They heard the main determiner form 60% of the time and 16 noiseforms that each occurred 2.5% of the time. The additional noise forms they heard (over and abovethose present in 8 ND condition) were /bIp/, /fu/, /z[1]l/, /zo/, /sep/, /mIb/, /lfm/, and /d[1]f/. This inputis represented graphically in Fig. 2, which shows the percentage of occurrence of each determinerform within each noun class across the different conditions. Each fill pattern represents one determinerform.As noted, the noise forms are less predictable than the main forms in two ways. First, they are lessfrequent. Second, they occur with nouns in both classes. This is not true of the main determiner forms.Note that the percentages given above are true within each noun class, but change when consideringFig. 2. Occurrence of determiner forms within each noun class in the input languages, across the five different conditions.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 9ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001nouns overall. In the 2 ND condition, for example, the percentage of all nouns occurring with each maindeterminer form is 30%, whereas the percentage of all nouns occurring with each noise form is 20%.While these overall percentages are much closer together than 60% and 20% (the percent occurrencesfor the main and noun determiners within each noun class), the main forms are still more frequentthan the noise forms even when considering nouns overall.In all other ways the input languages across the groups were as similar as possible. As mentioned,the noun phrases containing the main determiner forms were the same for all participants. Similarly,all noun phrases occurring with /te/ and /meg/ in the input of the 16 ND participants also occurred with/te/ and /meg/ in the input of the 8,4, and 2 ND groups. Similarly, a noun phrase containing /kum/ in the16 ND input also contained /kum/ in the 8 and 4 ND input sets. All other parts of the grammar were thesame, and completely consistent, in all four input groups.In order to be sure that the various determiner forms were actually inconsistent and probabilisticin their appearance (and not accidentally associated with a syntactic function or with a particular lexicalitem), the occurrence percentages for each determiner for noun phrases in general were alsomaintained for each syntactic position and for each noun. For example, in the presentation set ofthe 2 ND group, 60% of intransitive subjects, transitive subjects, and transitive objects occurred withmain determiner forms, and the 2 noise forms were similarly evenly distributed across each syntacticposition. While it was not possible to maintain precisely the same occurrence percentages for eachnoun, individual nouns occurred with particular determiner forms within a range centered aroundthe condition percentage. For example, in the 2 ND condition, the main forms occurred 41–74% ofthe time with individual nouns, with an average of 60% across nouns, and the noise determiners occurredwith particular nouns 9–33% of the time, with an average of 20%. In addition, each presentationof a particular sentence within and across sessions could be different from the other three. This lastpoint is particularly important, because it ensures that there were no consistent conditioning contexts.Thus there was no pattern of determiner use available to be learned from the input data, other thanthe percentages of use of the various forms.2.1.5. TestsParticipants were given four different types of tests to evaluate their performance. Tests were givenin the order in which they are described below.2.1.5.1. Vocabulary. A vocabulary test was given twice. The first was administered after participantswatched the videotape in the fourth session. In this task, participants were tested on their knowledgeof twelve vocabulary items. They were told that this test was designed to give them some idea of howthey were doing up to this point – that it was for their own benefit and would not be analyzed. Participantswere asked to provide a name for each object as it appeared on a video monitor and weregiven as much time as they needed to respond. All responses were videotaped, but (in accord withthe instructions) the results were not analyzed.A second vocabulary test was used to evaluate whether participants had learned enough vocabularyto be tested on more complex aspects of the language and was administered with the other testsin the final session. Participants were tested on the same 12 items as in the first vocabulary test, butthe order in which the items appeared was different. We tested the same nouns twice for one principalreason: these are the nouns required to complete the sentences in the production task, and we thereforewanted to direct attention to them in an implicit way. Post-test debriefing indicated that very fewparticipants had noticed this. Presentation and recording were the same as in the first vocabulary test.2.1.5.2. Sentence completion task. The test of primary interest was a sentence completion task. Thistask was designed to evaluate the participants’ production of determiners – the inconsistent part ofthe language. Participants saw a novel scene on the video monitor and heard the first word of the correspondingsentence. They were then asked to produce the complete sentence and were given asmuch time as they needed to provide an answer. For example, a participant sees a toy bird jumparound and hears the word /mErt/ ‘move’. She should then say /mErt ffmpogV po/ ‘move bird det’. Becausethe language is V-S-O, participants were always given the verb and had to produce the wholesentence, thus generating the NP(s) themselves.10 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001There were 24 test sentences (12 transitive and 12 intransitive), resulting in 36 possible NPs andtherefore 36 possible determiners. Participants were first tested on the transitive sentences and thenon the intransitives. The test set was designed so that 12 nouns each appeared once in each possiblesyntactic position (intransitive subject, transitive subject, and transitive object). The first use of theindividual noun varied between subject and object position in the transitive sentences; some nounswere first used as subjects and others as objects. Participants were asked to indicate where a nounthey could not recall should go in the sentence (for instance, by saying X instead of the noun). Thisallowed us to include the data from incomplete responses. Responses were videotaped and later transcribedfor analysis. All sentences used in this and other tests were novel to the participants and werenot part of the exposure set.2.1.5.3. Determiner judgment task. The third test was a grammaticality judgment task that also examinedparticipants’ knowledge of determiner usage, but through judgment rather than production. Participantswere asked to listen to 48 sentences one at a time and judge each of them on a four-pointscale according to how much they ‘liked’ or ‘disliked’ the sentence. Participants were instructed to respondthat they really liked a sentence when it sounded like a sentence from the language that theyhad been learning, and to respond that they really disliked a sentence when it sounded completelyunlike a sentence from the language. They were also told that if they thought a sentence was mostly,but not completely, like or unlike sentences from the language, they should use the middle of thescale. Participants responded by pointing to one of four different ‘happy’ or ‘sad’ faces. The experimentwas designed in this fashion so that it also could be done with children without changing the tasks.The 48 test sentences consisted of four variations of 12 base sentences: one form contained themain determiner form appropriate to the noun, one contained a noise determiner (one of the formsto which all subjects other than the control subjects had been exposed), one had the determiner inthe wrong location (preceding the noun), and one had no determiner at all. The sentences were randomlyordered, with the constraint that two versions of the same base sentence could not follow eachother. The four variations of one base sentence are shown in (2):(2) a. /gern ferluka po/ (correct: det follows noun)‘fall girl det’b. /gern ferluka meg/ (infrequent/incorrect: noise det)a‘fall girl det’c. /gern po ferluka/ (incorrect: main det precedes noun)‘fall det girl’d. /gern ferluka/ (incorrect/infrequent: no det)b‘fall girl’a This is a lower frequency form for all participants exposed to noise determiners, although the frequency varies by condition.For the control participants, however, it is an incorrect form which they have never heard.b As with example (b), the type of this example varies in its correctness by input group. For the control participants it is thelower frequency form in their exposure. For all other participants this type of sentence is an incorrect form to which they werenot exposed.Four of the 12 base sentences varied the determiner occurring with the transitive subject, four variedthe determiner occurring with the transitive object, and four were intransitive (and therefore variedthe determiner occurring with the subject). Sentences were presented on a SonyTM minidisk deckMDS-S38 through headphones, preventing the experimenter from hearing the sentence to whichthe participant was responding. This prevented the experimenter from being able to inadvertentlycue the participant to any particular response. Responses were recorded on an answer sheet by theexperimenter. Participants had 3 s to respond to each test item. Again, all sentences were novel.2.1.5.4. General Grammar Test. The fourth test, also a grammaticality judgment task, examined whatparticipants had learned about the rest of the language. Participants listened to 16 pairs of sentencesand were asked to select the sentence from each pair that sounded most like a sentence from theC.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 11ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001language that they had been learning. The two sentences in each pair were versions of the same sentence,one grammatical, the other ungrammatical. Test sentences were presented using the SonyTMminidisk deck MDS-S38 over headphones. Participants listened to both versions of the sentence andcircled 1 or 2 on an answer sheet, indicating whether they preferred the first or second sentence inthe pair. Half of the sentence pairs tested participants’ knowledge of verb subcategorization, that is,whether they knew that transitive verbs required two nouns and intransitives only one. The remainingsentence pairs tested whether participants knew that a verb was required in every sentence. Theserules of the grammar were tested for both transitive and intransitive sentences. For the transitive sentenceswith missing arguments, either the subject or the object could be the missing argument. Whichsentence (first or second) in the pair was grammatical was randomized, as was the ordering of sentencepairs in the test, with the constraint that no more than two sentences could occur in a row thattested the same rule and were of the same valence. There was a 1-s pause between the two sentencesthat formed a pair and a 5-s pause between pairs. Pairs were not identified as such, except by theoccurrence of the longer pause. Again, all test sentences were novel; none appeared in the exposureset.2.2. ResultsWe begin by reporting results for tests that demonstrate more general knowledge of the language,and then move on to describe the results on determiners that are of primary interest.2.2.1. Vocabulary testsIn accord with the instructions given to participants, the data from the first vocabulary test werenot tabulated. On the second vocabulary test, all participants scored at least 5 out of a possible 12(the criterion we have previously used for deciding whether participants would be given certain ofthe remaining tests), with a mean of 8.88 items (SD = 2.39) for participants overall. Means variedslightly across the input groups, ranging from 7.83 (16 ND) to 10 (2 ND). A one-way ANOVA with inputcondition as a between-subjects factor indicated that the differences in vocabulary scores were notsignificant.2.2.2. General Grammar TestThis test examined participants’ knowledge of parts of the grammar other than determiners. Weconducted this test to ensure that learners in all conditions successfully acquired those parts of thegrammar represented consistently in the input. The test examined participants’ knowledge of sentenceconstruction (did they know that a verb is required in every sentence) and verb subcategorization(did they know that one set of verbs is transitive, requiring two nouns, and another intransitive,requiring only one noun).Fig. 3 shows the mean score by noise condition. The overall mean was 14.5 out of a possible 16,SD = 1.66. This was significantly and substantially above chance (t(59) = 30.29, p < .001). We conductedFig. 3. Mean score on general grammar test by input group. (In this and all other figures, error bars represent standard error.)12 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001a repeated measures ANOVA with rule type (two levels) and transitivity (two levels) as within-subjectfactors and noise condition (five levels) as a between-subjects factor. The main effect for noise conditionwas not significant, indicating that the manipulation of determiners had no effect on participants’learning of the other, consistent parts of the grammar. There was a significant effect of rule type(F(1,55) = 9.2, p = .004), with participants scoring slightly higher on knowledge of basic sentencestructure (every sentence must have a verb) than on verb subcategorization (7.55 vs. 6.95 out of apossible 8) – not surprising since the former requires very general knowledge of the language, whilethe latter depends on knowledge of particular verbs. There was also a significant main effect of transitivity(F(1,55) = 16.55, p < .001), with participants performing slightly better on transitive test itemsthan intransitive ones (7.63 vs. 6.87 out of 8). There were no significant interactions.Overall, participants performed very well on this test, indicating that they had learned these consistentfacets of the grammar. Moreover, their performance was not affected by the amount of inconsistencyof the determiners: participants performed equally well in all input conditions.2.2.3. Sentence completion taskThe results of this test were of primary interest. It permitted us to observe the effect of the presenceand amount of scatter in the linguistic input on the production of determiners. In particular, wewanted to know whether, when exposed to scattered inconsistency in determiner usage, participantswould reproduce the inconsistency present in their input or would regularize the more frequent formsto which they were exposed. For each participant we computed the percentage of main determinerproduction (the number of correct main determiners used by the participant, divided by the numberof possible determiner usages, multiplied by 100). The number of possible determiner usages was simplythe number of correct nouns produced by the participant in this task. Fig. 4 shows the mean percentageof main determiner production for the five input groups. Recall that the percentage of mainforms in the input was the same – 60% – for all input groups.As can be seen in Fig. 4, participants exposed to scattered inconsistency (input containing noisedeterminers – all of the ND conditions) produced more main determiner forms than those exposedto presence/absence inconsistency (main determiner forms alternating with determiner omission –the Control condition). A one-way ANOVA confirms that there is a significant effect of input conditionon the production of determiners (F(4,55) = 8.27, p < .001), and a t-test comparing the control groupwith the four scatter groups is significant (t(55) = [1]4.889, p < .001, with pooled variance estimate).4Fig. 4. Mean production of main determiner forms by input group.4 Tests for homogeneity of variance were not significant, indicating that it is licit to pool the variances. The contrast is alsosignificant using a separate variance estimate (p<.001).C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 13ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001We were particularly interested in whether increasing the noise, or scatter, would induce increasedregularization behavior in adult learners. The data indicate that this is indeed the case: As participantsare exposed to increasing numbers of noise determiners, they produce increasing percentages of maindeterminer forms (Flinear(1,57) = 25.92, p < .001).5 This effect shows that extensive scattered inconsistencydoes produce regularization, with participants in the 16 ND producing almost 90% main determinerforms, even though their input contained these forms only 60% of the time.The design of the test allowed us to examine whether participants were using determiner formsdifferently according to the syntactic position of the noun (subject of an intransitive verb, subject ofa transitive verb, or object of a transitive verb). A repeated-measures ANOVA with input conditionas a between-subjects factor and syntactic position as a within-subject factor showed no effect of syntacticposition and no significant interaction between input and syntactic position. This indicates thatparticipants were not imposing on the input a more deterministic, linguistically-based rule, such asusing the main determiners in association with subjects versus objects or with transitives versusintransitives.What were participants doing when not using the main determiner forms? Control participantsused no determiner form at all. Thus they were replicating their input, as we found in our earlierstudy. Participants in the noise determiner input groups primarily used noise forms, with many participantsusing 1 or 2 noise forms to the exclusion of the others. (This is the only possibility in the 2 NDcondition. However, this was common among participants in the 8 and 16 ND groups as well.) This useof a few noise determiners did not reduce the amount of inconsistency present in the language, however,as different speakers preferred different noise forms. Most participants preferred the phonologicallysimpler CV forms, but this trend did not hold for all participants; some used a CVC form as theirpreferred noise form. There were also some who used the incorrect main form (the main determinerform for the other noun class), and some who created a novel determiner form (usually blends of oneor more existing forms). Occasionally, a participant in a scatter condition used a bare noun, somethingnot present in their input. Thus, participants were not regularizing the noise or scatter in their input;their non-main form productions remained inconsistent and noisy. Regularizations, when they occurred,involved more frequent use of the main determiner forms.In sum, speakers in the control group basically reproduced their input. In contrast, participants innoise groups showed a tendency to regularize their input, using the most frequent forms more oftenthan they had heard them. Moreover, increased complexity of variation in the input resulted in increasedregularization, with participants producing increasingly more main forms in their speech asthe number of noise forms in their input increased. Participants exposed to a few noise forms producedthe main determiner forms only slightly more often than they had heard them, though muchmore often than participants who heard no noise forms at all (presence/absence inconsistency); participantsexposed to 16 noise determiners produced the main determiner forms almost 90% of thetime, a full 30% more often than they heard them.2.2.4. Determiner judgment taskThis task was designed to assess participants’ knowledge of determiners in a different way –through grammaticality judgments. As described above, participants were asked to rate 48 novel sentences,one at a time. Twelve of the test sentences were correct, 12 contained a noise determiner (towhich all subjects other than the control subjects had been exposed), 12 had the determiner in thewrong location (preceding the noun), and 12 had no determiner at all.Fig. 5 shows the mean ratings given by participants to each kind of sentence for the five inputgroups. A MANOVA with sentence type and syntactic position as within-subject repeated-measuresvariables and input groups as a between-subjects variable was conducted on the data.5 This F was computed adjusting for unequal spacing between the categories of the factor (input condition). Because it is notclear whether the control condition falls along a continuum with the scatter conditions, we also conducted the linear trend analysiswith only the 4 noise conditions included, and it too is significant (F(1,45) = 11.21, p = .002).14 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001The primary variables of interest are the effects of input group and determiner manipulation. Themain effect of input group is not significant. The main effect of determiner manipulation is significant(F(2.4, 130.9) = 368.31, p < .001).6 This is modulated by a significant interaction between the two variables(F(9.5, 130.9) = 35.29, p < .001), reflecting the fact that all participants liked sentences with main determinersand disliked sentences with the determiner in the wrong location, but the groups differed in theirratings of sentences without determiners and with noise determiners. We were particularly interestedin the ratings given by participants to sentences which they had encountered in their input. For the controlgroup these were sentences with main determiners and those without determiners. For all othergroups these were sentences with main determiners and those with noise determiners. For each inputgroup, we compared the ratings given to these two sentence types. The results show that all participantsreliably rated the more frequent sentences higher than the less frequent sentences. Repeated-measuresANOVAs with sentence type as a within-subjects repeated-measure were significant for all five inputgroups: control group (F(1,11) = 12.19, p = .005); 2-noise (F(1,11) = 7.79, p = .018); 4-noise (F(1,11) = 20.51,p = .001); 8-noise (F(1,11) = 12.20, p = .005); 16-noise (F(1,11) = 60.71, p < .001).We also examined the data to see whether the degree of difference between the ratings given to thetwo types of sentences increased as the number of noise determiners increased. That is, did the participantsexposed to 16 noise determiners distinguish between sentences with main determiner formsand those with noise determiner forms to a greater degree than participants exposed to fewer (or no)noise forms? To examine this, we computed a difference score for each participant between the meanrating given to sentences with main determiners and the mean rating given to whichever kind of sentencewas the other one in her input (sentences with no determiners for control participants and sentenceswith noise determiners for all others). We then performed a trend analysis on the differencescores. This analysis is significant (F(1,56) = 22.89, p < .001), indicating that the difference in ratingsdoes indeed increase as the number of noise determiners increases, mirroring the trend we foundin the production data.Also as in the production task, we asked whether participants would judge determiner forms differentlyaccording to the syntactic position of the noun, indicating that they might be imposing on theinput a more deterministic linguistic rule, such as using the determiner forms in association with subjectsversus objects or with transitive but not intransitive sentences. Such a tendency would be reflectedin a significant interaction between syntactic position and sentence type. However, this wasnot the case. Although the main effect of syntactic position is significant (F(2,110) = 4.59, p = .012), there6 The degrees of freedom used in the sentence/determiner manipulation analyses are adjusted using the Huynh-Feldt Epsilondue to a significant test for heterogeneity of variance.Fig. 5. Mean ratings by input group.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 15ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001is not a significant interaction between sentence type and syntactic position, nor is there a significantthree-way interaction between sentence type, syntactic position, and input group.In sum, to some degree participants’ judgments reflected their input. Sentences with the determinerin the wrong location were universally disliked by all groups. None of these sentences occurredin the input of any grou#p#分页标题#e#p. Likewise, all participants preferred the sentences with the main determinerforms, which were the most common sentence form in the input for all participants. Where participants’ratings differed was in the ratings given to sentences they had heard, but less frequently (thatis, those with noise determiner forms or those without determiners, depending on condition). Participantsrated those sentences they had heard less frequently substantially higher than those they hadnever heard. Importantly, however, participants also showed the same regularizing tendency in theirjudgments as was reflected in their productions: with increasing numbers of noise determiners, theyshowed an increasing tendency to prefer the main determiner forms over less frequent forms.2.3. DiscussionThe data from Experiment 1 clearly demonstrate that regularization behavior can be induced inadult language learners when they are given input that contains what we have called ‘scattered inconsistency.’This contrasts with the results from our earlier work (Hudson Kam & Newport, 2005), whereadult participants given input containing presence/absence inconsistency did not regularize or overusethe main determiner forms, but instead used these forms with almost exactly the probabilitieswith which they appeared in the input. Interestingly, this parallels the findings seen in natural languageacquisition in children like Simon, who regularized the use of his parents’ most frequently usedmorphemes (Singleton & Newport, 2004). Of note is that Simon’s parents did not show a simple alternationbetween the presence and absence of required ASL morphemes, but instead showed variationof forms that was more like scattered inconsistency: they used the correct form most of the time, butwith low and variable frequency might replace this form with any of several different incorrect forms.It also parallels the hints of similar effects seen in probability learning, where adults usually probability-matched, but when asked to make predictions over more than two lights, often displayed overmatching(Gardner, 1957; Weir, 1964, 1972). This similarity suggests that perhaps the samemechanisms are at work in response to inconsistencies in natural language acquisition, in the languagelearning modeled in this paper, and in the kind of learning investigated in basic probabilitylearningexperiments.However, a number of questions remain regarding this phenomenon, and especially how and whywe have been able to induce regularization in adult learners. It is important to note that extensive regularizationoccurred in this experiment only in the extreme case where there were 16 noise determiners,each appearing only 2.5% of the time, varying with main determiners that appeared 60% of thetime. In subsequent experiments we will address two important questions about this finding. First,perhaps this apparent regularization occurred only because of the low frequency of the noise determiners,and not because of the inconsistency of main and noise determiner use that characterizesthe natural language phenomena with which we began our studies. To address this question, we willcompare performance in the 16 ND condition of Experiment 1 with a frequency-matched but differentlystructured condition in Experiment 2, in which the same number of determiners are used withthe same overall frequencies, but where their appearance is perfectly regular and consistent. Subsequently,in Experiment 3 we will investigate learning with scattered inconsistency in children as comparedwith adults, to see whether scatter has the same effects in children, or rather whether childlearners show tendencies to regularize that are more independent of the nature and extent of theinconsistency than is the case for adults.3. Experiment 2The purpose of Experiment 2 was to assess whether the regularization found in adult learners inthe 16 ND condition of Experiment 1 was due to the scattered inconsistency of the noise determinersas compared with the main determiners, as we hypothesized, or rather whether it resulted more sim-16 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001ply from the difficulty of learning any forms that occur with the low frequency of these noise determiners.In the present experiment, two main determiner forms occurred 60% of the time and 16 lowerfrequency determiner forms occurred 2.5% each, but their appearance was strictly conditioned by theoccurrence of particular nouns with which they were associated. The question of interest was whetheradult learners also regularized the main determiners under these circumstances, or rather whetherthey were able to learn the low frequency determiners as well as the main determiners when eachoccurred in structured contexts.3.1. Method3.1.1. ParticipantsEleven adults, mean age 20.9 years, participated in this study. All were students at the University ofCalifornia, Berkeley at the time of the study. Participants were recruited through flyers posted on campus.They were paid daily for their participation and received a bonus upon completion of theexperiment.3.1.2. The LanguageWe used the same basic language as in the 16 ND condition of Experiment 1, with one importantdifference. In this study, though the two main determiners and the 16 low frequency determiners occurredwith the same frequencies as in Experiment 1 (that is, in 60% of the noun phrases for each of thetwo main determiners and in 2.5% of the noun phrases for each of the 16 low frequency determiners),their appearance was perfectly regular and consistent. To achieve this consistency, each of the determinerswas assigned to particular nouns and occurred every time these nouns occurred (that is, thedeterminers were lexically consistent). However, by varying the number of nouns assigned to eachdeterminer, we could create the same high and low frequencies for the determiners as was the casefor the 16 ND condition. Nouns were divided into 18 arbitrary classes, 2 large and 16 small. One largeclass contained 11 nouns, the other contained 9. The small classes each contained a single noun. Asbefore, there were no differences in meaning or phonology between the nouns in different classes;the only grammatical consequence of noun class was determiner selection: each class of nouns tooka different determiner.3.1.3. PresentationPresentation was the same as in Experiment 1. As before, exposure and testing were conductedindividually for each participant.3.1.4. Experimental manipulationAs mentioned above, nouns were divided into 18 classes, 2 large classes and 16 small ones containinga single noun, with each class taking a different determiner. This particular division allowed us topresent input sentences that were exactly the same as in the 16 ND condition of Experiment 1, exceptwith respect to which determiner occurred with which noun phrases. Moreover, the input set containedthe same overall distribution of determiners as in the 16 ND condition. That is, the probabilityof any individual determiner given a noun (any noun) was almost exactly the same in the two experiments(p DETi|noun in Experiment 2 = p DETi|noun in 16 ND condition).7In terms of the number and overall distribution of the determiners, the languages are equally complex.Importantly, however, the occurrence of the determiner with any particular noun is more consistentand predictable in the present language (p = 1) than it was in the previous experiment(p = 0–0.722). All other aspects of the grammars were the same, and completely consistent, in thetwo experiments.7 To be precise, the percentages for each determiner differed by .1 or .2% for 8 of the low frequency determiners (and otherwisewere exactly the same) and differed by 1% for one of the high frequency determiners. These tiny differences between Experiment 1and 2 arose from using the same input set and tests for the two experiments and are extremely unlikely to be responsible for anydifferences in results.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 17ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.0013.1.5. TestsTo evaluate what they had learned about the language, participants were given three of the fivetests used in Experiment 1: a vocabulary test, the sentence completion task, and the forced-choicegeneral grammar test. The sentence completion task contained five large class nouns and seven smallclass nouns. Tests were constructed and presented as in Experiment 1. All tests were administered inthe final ninth session.3.2. ResultsRecall that the language to which we exposed participants in this experiment contained the sameoverall distribution of determiners as the condition in Experiment 1 in which participants regularizedmost (the 16 ND condition). In many ways, then, the two languages are equally complex, though one iscompletely regular and the other contains scattered inconsistency. To assess the degree to which complexityalone leads to particular learning outcomes, and to assess whether regularization of the dominantdeterminers might be due purely to the low frequency of the noise determiners, we comparedthe data from Experiment 2 with those from the 16 ND condition in Experiment 1. We begin with thegeneral tests and proceed to the ones of principal interest. Note that in the following analyses, wheneverthere is a reference to Experiment 1 it is to the 16 ND condition only.3.2.1. Vocabulary testParticipants in this experiment knew an average of 9.82 (SD = 2.04) vocabulary items (out of 12).The minimum score was 7, the maximum was 12. The vocabulary score for participants in Experiment1 was slightly lower at 7.83 (SD = 1.9), a significant difference (F(1,21) = 6.782, p = .017). However, thehigher vocabulary performance for participants in Experiment 2 is within the range of conditionmeans from Experiment 1.3.2.2. General Grammar TestParticipants performed very well on this test, indicating that they did indeed learn the language.The mean scores for the two groups of participants were 13.82 (Exp. 2, SD = 2.79) and 14.12 (Exp. 1,SD = 1.47) out of a possible 16. We subjected the data to repeated-measures ANOVA with rule type(two levels) and transitivity (two levels) as within-subject factors and experiment (two levels) as a between-subjects factor. The main effect of experiment was not significant, indicating that the twogroups of participants learned the language equally well. The overall mean score of 14.00 out of a possible16 (SD = 2.15) was significantly above chance (two-tailed t(22) = 13.36, p < .001). As in Experiment1, transitivity was significant (F(1,21) = 5.32, p = .031): participants performed slightly better on testitems involving transitive sentences than those involving intransitives (7.43/8 vs. 6.57/8). Also as inExperiment 1, none of the interactions were significant. Unlike Experiment 1, rule type was not significant;participants performed equally well on items testing both rules.3.2.3. Sentence completion taskThis test examined participants’ use of determiners. In particular, we were interested in whetherparticipants exposed to a complex but consistent pattern of determiners would regularize the languagelike participants in Experiment 1. Fig. 6 shows the mean percentage of correct determiner usesfor nouns from the large and small classes. Participants overwhelmingly used the determiner formthat was correct for the noun. This is true for nouns from both the large and small classes, althoughparticipants made fewer determiner errors with nouns in the large classes (t(10) = 2.39, p = .038).8While these results suggest that participants in Experiment 2 are not regularizing or over-using themain determiners, it is necessary to score the results in a different way to see this more clearly. In orderto compare the productions of participants in the two experiments, we need a metric that is comparableand equally meaningful for each of the conditions. Overall production of large and small class8 The figure and related analyses are based on 11 nouns, not 12, for participants in this experiment only, because all participantsconsistently got one of the small class nouns wrong (and therefore could not be scored on usage of its determiner).18 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001determiners versus main and noise determiners is not such a measure. While the nouns do not matterfor participants in Experiment 1, since the proportion of each kind of determiner produced or observedin the input does not depend on them, they do for participants in Experiment 2, since determinerchoice depends crucially on the identity of the noun. To obtain a score that is comparable for thesetwo different circumstances, we computed two proportion scores for each participant, one for mainor large class determiners and the other for small class or noise determiners. Each score reflects theproportion, given what would be expected from the input, of correct determiner production for thattype of determiner. A proportion of 1 represents exactly matching the input, while anything aboveor below 1 is a deviation from the input.9 What constitutes ‘correct’ is different for participants inthe two experiments. For participants in Experiment 1, correct responding is probability matching. Forparticipants in Experiment 2, correct responding is getting the determiner correct for the specific noun.For example, if a participant in Experiment 1 produced 10 nouns, 8 with the correct main determiner and2 with a noise determiner, the proportion of main determiners produced would be 1.33, since they areproducing a third more main determiners than expected, and the proportion of noise determiners producedwould be .5, one half of the four that would be expected were the participant matching theprobabilities.Fig. 7 shows the mean proportions on this measure for participants in the two experiments, forlarge class or main determiners and small class or noise determiners, with correct responding (1.0)indicated by a dotted line. It is clear from the figure that participants in Experiment 1 are producingfar more main determiners than they heard and far fewer noise determiners. Participants in Experiment2, by contrast, are slightly under-producing both types of determiners (large and small class)but otherwise approximately producing the language that they were exposed to. The difference betweenthe two groups is significant for both types of determiners (Main/Large class: F(1,21) = 41.67,p < .001; Noise/Small class: F(1,21) = 53.75, p < .001).10 Thus, despite hearing the same proportions of18 different determiners, participants in Experiment 2 are much better able to reproduce what they’veheard.A complementary view of this difference is provided by examining what participants are doingwhen they do not match their input. The relevant productions for participants in Experiment 2 are er-9 This measure corrects for the fact that different participants produced different nouns, and thus each had different underlyingprobabilities of production. Let’s imagine two participants who both produced six nouns. One produced four large class and twosmall class nouns, and failed to produce any small class determiners, the other produced two large class and four small class nouns,and produced half of the small class nouns with the correct determiners. Both would have undershot the target percentage by33.33%. However, one is getting rid of the small classes, the other is not. Using a proportional measure captures this.10 These two measures are not necessarily reciprocal because these are not the only possible types of productions.Fig. 6. Mean production of correct determiner forms for nouns in large and small classes.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 19ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001rors. For participants in Experiment 1, ‘errors’ include overproduction of the noise and main forms, aswell as true errors. As already noted, participants in Experiment 2 had far fewer errors; on average,only 16.3% of their noun phrases fell into this category (SD = 4.4), compared to 44.12% (SD = 3.02) inExperiment 1. Here we examine the proportions of the different types of errors participants made; despitetheir low rate of errors, it is possible that participants in Experiment 2 might have shown someregularization, for example, preferentially extending large class determiners when they did makeerrors.Fig. 8 shows these errors for the two experiments, divided into four types: (1) overuse of the correctmain determiner (only possible for Exp.1, not possible for Exp. 2), (2) use of the incorrect main determiner(possible in both experiments), (3) incorrect noise forms (only possible in Exp 2), and (4) zeroforms (determiner omission). The figure clearly shows that the participants in Experiment 2 did notregularize like participants in Experiment 1. While they did make errors, these were few, and impor-Fig. 7. Mean over- and under-production of main/large class and noise/small class determiners – Experiments 1 and 2.Fig. 8. Mean proportion of ‘errors’ of different types – Experiments 1 and 2.20 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001tantly did not involve preferentially overextending the more frequent large class determiners. For participantsin Experiment 2, the percentage of errors that were incorrect uses of the large class determinerswas not significantly different from incorrect uses of the less frequent small class determiners orfrom using no determiner at all.3.3. DiscussionOverall, these results indicate that distributional complexity per se is not enough to cause adultlearners to regularize a grammatical form. The results also suggest that it is not merely the low frequencyof the noise determiners that produced regularization in Experiment 1. Rather, it appears thatinconsistency in how a form is used, in combination with inconsistency as well as low frequency inhow competing forms are used (what we have called scattered inconsistency), are necessary for adultlearners to regularize and extend the form’s usage.The question remains, however, whether the regularization seen in adult learners of Experiment 1demonstrates the same mechanism involved in cases of actual language change, particularly creolization(see Hudson Kam & Newport, 2005, for a discussion of the hypothesis that adult learners might beresponsible for regularization in creole languages). Although the participants exposed to scatteredinconsistency in Experiment 1 did produce the main determiner forms more often than they hadheard them, they did not, in most of the noise conditions, fully regularize the inconsistent forms (thatis, use them virtually all the time in a given context). Only the participants exposed to 16 noise determinersapproached truly rule-like usage of the main determiner forms, using them close to 90% of thetime. Although pidgins and incipient creoles may exhibit scattered inconsistency like that modeledhere, they do not, as a general rule, contain nearly so many forms in competition with one another,even across speakers. In the Tok Pisin example mentioned earlier, for instance, there were five formsin competition for marking continuous aspect (Sankoff, 1979); but the adult learners in Experiment 1who were exposed to this amount of scattered inconsistency produced the main determiners onlyslightly more than they heard them in the input. Given these results, it does not seem likely that adultlearners of a pidgin or incipient creole would regularize one of these forms to the degree that would benecessary to explain the rapid linguistic change hypothesized to occur in creolization. This in turn suggeststhat, although adults can regularize under certain circumstances, it is unlikely that they are theprimary agents of regularization in the circumstances of real language change. In the next experiment,we investigate this question further, by directly comparing child learners with adult learners, to seewhether children regularize more readily than adults and do so under circumstances closer to thoseof natural language change.4. Experiment 3In previous work we found that children regularized inconsistencies that adults reproduced, suggestingthat children might be more likely than adults to regularize (Hudson Kam & Newport,2005). Work by Newport and colleagues on children learning American Sign Language from non-nativeinput also found that children regularize more than adults; the children they studied regularizedto a much higher degree than most of the adults in Experiment 1 (Newport, 1999; Singleton & Newport,2004, Ross & Newport, in prep). The latter differences in outcome could be due to the differencein language modality, but a more likely possibility is that it is due to a difference in the ages of thelearners: children might regularize more readily than adults across a variety of circumstances. Experiment3 explicitly compares children and adults when learning several different types of inconsistentlanguages. In particular, we ask whether children and adults differ or look the same in learning versusregularizing inconsistencies, examining both presence/absence inconsistencies and scattered inconsistencies.In addition, we ask, when adults and children do regularize inconsistency, whether theydo so in the same way (though perhaps to a different degree), or rather whether they seem to be performingdifferent processes.Importantly, in this study we compare children’s and adults’ learning of the type of inconsistencythat we know (from Experiment 1) adults regularize, what we have called scattered inconsistency.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 21ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001However, we introduce only the lower levels of scatter – the 2 ND and 4 ND conditions – which the adultparticipants of Experiment 1 did not regularize very much. This allows us to see if children are morelikely than adults to be systematic, perhaps doing so at lower degrees of scatter than adults, orwhether they regularize under different circumstances than adults. Our own previous work suggeststhat this is a possibility. In Hudson Kam and Newport (2005), we found that children will regularizepresence/absence inconsistency, a type of inconsistency that adults do not regularize but insteadreproduce quite accurately in their speech.We exposed children ages 5–7 to languages containing presence/absence inconsistency and somelimited scattered inconsistency and then tested them to see what they had learned about the consistentand inconsistent facets of the language. In order to allow the languages to be mastered by children,the methods used in Experiment 3 were simplified from those in Experiment 1 in severalways and so are described in some detail. Because of these differences, we also tested a small numberof adults with the same procedures, to ensure that any differences in results we found between adultand child learners are actually due to the age differences and not to differences in methods.4.1. Method4.1.1. ParticipantsForty children participated in the study. Of those, 30 completed the study. Three children stoppedattending child care in the middle of the study (two children left to go on holidays, one got sick) andseven children either did not know enough nouns to complete the production task or could not produceany sentences and so did not complete the study. Mean age of the 30 children was 5 years,10.6 months. Thirteen of the children were male, 17 were female. Sixteen adults participated. Theyhad a mean age of 20 years, 2.25 months. Two of the adult participants were male, 14 were female.Child participants were recruited through local daycares and preschools that had agreed to participate.Parental consent was first obtained, and then each child was asked whether they would personallyassent to participate. Most received a small toy at the end of each session. (This was against thepolicy of one of the preschools.)All adult participants were students at the University of Rochester or University of California,Berkeley, at the time of the study. Adult participants were recruited through posters (Berkeley) oremails sent to people in the department subject pool (Rochester) describing the study and invitingthem to participate. All were paid daily for their participation and received a bonus upon completionof the entire experiment.4.1.2. The LanguageThe basic language contained 17 words: 4 verbs, 12 nouns, and 1 determiner. Unlike the languagein Experiment 1, there was only one noun class and therefore only one main determiner. (The vocabularywith glosses is given in Hudson Kam & Newport, 2005). Although this is a larger vocabulary thanis often used with children in artificial language experiments (cf. Moeser & Olson, 1974, but see alsoBraine et al., 1990 and MacWhinney, 1983), it was learnable to some extent by almost all of the children.The lexicon, in conjunction with the objects used and the constraints they impose, result in 99semantically possible sentences. Participants were exposed to only a sample of these sentences; therest were reserved for testing.4.1.3. PresentationParticipants were typically run in groups of two or three as they were available to us. Givenchanges in the daily availability of the children, the grouping of the children could change from dayto day. This allowed us to run numerous children at the same site within as short a time as possible.Adult participants were also run in groups, but always of two and always with the same partner. However,as described below, all testing was done with participants individually.The exposure set consisted of 12 intransitive sentences and 12 transitive sentences. Each of thetwelve nouns in the language appeared once in each syntactic position (intransitive subject, transitivesubject, transitive object) in the exposure set. The intransitive sentences were split equally betweenthe two intransitive verbs: six sentences were ‘fall’ events, six were ‘move’ events. The transitive sen-22 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001tences consisted of three ‘inside-of’ events and nine ‘hit’ events. This reflects the fact that there aremore possible ‘hit’ events than there are ‘inside-of’ events.Pilot work suggested that the videotaped exposure method used in Experiments 1 and 2 was ineffectivefor use with children, so in this experiment we used live exposure. Children also found it difficultto learn the vocabulary from their presentation in sentences, so we directly taught participantsthe vocabulary items as well as their meanings. However, there was no explicit teaching of the grammaticalaspects of the language; as in the previous experiments, participants acquired the grammarentirely through exposure to sentences and their accompanying events.11 Importantly, the same methodswere used with the adult participants in this experiment as were used with the child participants.There were six exposure sessions, each of which lasted approximately 10–20 min. The seventh sessionwas a test session. The seven sessions were completed over nine days by all participants.Exposure proceeded as follows: The experimenter began by explaining to the participants that shewas going to teach them a new language called Sillyspeak; first, they would learn some new words forthings, and then some new ways to say things. For the adults, exposure began at this point. For thechildren, the experimenter would often chat for a few moments with the children, explaining whatit means to learn another language. Most children did not know, or at least did not really understand,the words ‘word’, ‘sentence’, or ‘language’, so the concept of other languages often revolved aroundother people talking in ways that they did not understand. (Many of them had grandparents or parentswho spoke other languages, and one day-care was teaching the children signs.). After this chat, exposurebegan.On the first day participants were taught the vocabulary, excluding the determiners. The entire listwas run through four times. Each run through the vocabulary began with the four verbs, always in thesame order. The experimenter would say ‘‘if you want to say ‘hit’ in Sillyspeak you say /flIm/,” then thesame thing for /prAg/ ‘inside of’, /mErt/ ‘move’ and /gern/ ‘fall’. Participants were asked to repeat theSillyspeak word after they heard it. Each of the verbs was accompanied by a gesture, and participants(especially the children) often also repeated the gestures, although they were not asked to do so. Afterrunning through the verbs, they were taught the nouns. On the first run through the noun vocabulary,participants were shown a toy and asked to name it, and were corrected if required. This was done toensure that they were encoding the intended meaning. The experimenter then told participants howto say the word in Sillyspeak. This was repeated three more times without having the participantname the object. The nouns were presented in a randomized order each time that was the same forall participants.Sentences were first introduced in the second session. This session began by going through thevocabulary list once. The experimenter then demonstrated how to ‘put words together’ to ‘say biggerthings.’ The experimenter began with the 12 intransitive sentences. She showed the participants anevent involving the toys and then said the corresponding sentence out loud (read from a piece of paperon her lap). Participants were asked to repeat the sentence after hearing it. After the intransitive sentences,the experimenter went through the vocabulary a second time, and then went on to the 12 transitivesentences in the exposure set, done the same way as the intransitives. The exposure sentenceswere always performed in the same order. The third and fourth days proceeded in exactly the sameway: vocabulary, intransitive sentences, vocabulary, transitive sentences. Day five consisted of onepass through the vocabulary, then the intransitive sentences, then the transitives, and then the intransitivesagain. Day six consisted of one pass though the vocabulary, then the transitive sentences, theintransitive sentences, and finally a second pass through the transitives. This design allowed 12 passesthrough the vocabulary and six through each kind of sentence. Participants were allowed to help theexperimenter act out the sentences to maintain their interest and attention.Occasionally participants had difficulty repeating a sentence. When this happened, the experimentersaid the sentence a second time. This happened almost exclusively with the children andwas most common on days 2 and 3, the first and second times they heard the sentences. When it oc-11 The children did frequently ask questions about the language, such as what the function of the determiner was. Theexperimenter told the children that she did not speak the language, however, and so could never answer their questions. Theyseemed to believe this, frequently making comments to each other about how the experimenter had to read the words off of apiece of paper, as if to confirm that indeed, she did not speak the language and so really could not answer their questions.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 23ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001curred later, it was usually due to inattention: the children were often run in rooms where other childrenwere playing and sometimes were distracted by their activity.4.1.4. Experimental ManipulationAll participants were exposed to the same basic sentences; input sentences differed across conditionsonly in the use of determiners. There were four determiner conditions in this experiment: completelyconsistent use of the determiner (100%), 60% presence/40% absence of the determiner (0 ND),60% occurrence of the main determiner form, alternating with 2 noise determiners that each occurred20% of the time (2 ND), and 60% occurrence of the main form and 4 noise determiners, each occurring10% of the time (4 ND). As in Experiment 1, these percentages were true for nouns overall; individualnouns occurred with the various determiners within a range that averaged to 60%, 20%, and 10% asappropriate. Note that these percentages were true only of the nouns occurring within sentences. Duringvocabulary training, nouns were presented without determiners. This was the same for participantsin all four input conditions.4.1.5. TestsIn the final session participants were given three different tests to evaluate their performance. Twotests were given to evaluate participants’ knowledge of determiners; one test examined their knowledgeof the consistent aspects of the grammar. Testing was always done individually. Tests were givenin the order in which they are described below.4.1.5.1. Sentence completion task. This task was designed to elicit the production of noun phrases, thepart of the sentence containing the inconsistency, in order to evaluate whether participants’ determinerproductions varied with the type of inconsistency present in the input.As in Experiment 1, we used a sentence completion task to accomplish this. First, the participantwas shown a series of toys and asked to name them. This continued until she had named five to sevenobjects or it became clear that she did not know any more nouns, whichever came first. Objects whichhad been named became part of the participant’s test set. Objects were selected (for showing) in twoways. First, the participant was always shown at least two of the three container objects (cup, barrel,and truck), since only these objects can be used with the verb /prAg/ ‘inside of.’ Second, toys that hadbeen remembered by previous participants were shown early. Often the participant would begin tospontaneously produce words she knew (both children and adults did this), and when this happenedthe experimenter asked what the word meant. If the participant produced the correct English word orretrieved the correct object, the object was included in the test set.Once a set of objects that the participant could name had been selected, the sentence completiontask began. Using the objects that the participant had named, the participant was shown an event orscene and was told what the sentence should mean in English, and was told what the first word of thecorresponding Sillyspeak sentence was. For example, the experimenter would wind the bear up(which made it move) and put it down in front of the participant and say something like, ‘‘OK, I wantyou to tell me how to say ‘the bear moves’ in Sillyspeak. The first word would be /mErt/, right?” If theparticipant had difficulty, they were reminded that they knew how to say things like ‘the bear falls’and ‘the rhinoceros moves’ in Sillyspeak, but they were not reminded how to say these familiar sentencesin Sillyspeak. The only Sillyspeak they were given by the experimenter was the relevant verb.Many children expressed a lack of confidence with the transitive (long) sentences, so we alwaysbegan with the intransitive (short) sentences. This allowed the children to gain confidence with thetask before attempting the longer transitive sentences. Transitive and intransitive sentences wereinterspersed. The experimenter wrote down each response before moving on to the next sentence.A subset of participants was videotaped and their productions later coded for reliability by a secondcoder who was blind to experimental condition. As in all tests, the test sentences were novel; they hadnot occurred in the exposure set.4.1.5.2. Determiner judgment task. The second test administered was a grammaticality judgment taskthat also examined participants’ knowledge of determiner usage. Participants listened to 18 sentencesand judged each of them on a four-point scale, according to how much they ‘liked’ or ‘disliked’ the sen-24 C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxxARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001tence, by pointing to one of four faces ranging from happy to sad. Participants were instructed to respondthat they really liked a sentence (picking a happy face) when it sounded like a Sillyspeak sentence,and to respond that they really disliked a sentence (picking a sad face) when it soundedcompletely different from Sillyspeak. They were also told that if they thought a sentence was mostly,but not completely, like or unlike sentences from the language, they should use the middle of the scale(slightly happy and slightly sad faces).The 18 test sentences consisted of three variations on six base sentences. The form of the variationsdiffered for participants in different input conditions. For all participants, one form of the sentencewas correct, and one had no determiner at all. For participants who had heard either completely consistentinput or input containing presence/absence inconsistency, the third variant had the determinerin the wrong location (preceding the noun). For participants whose input had included noise determiners,the third variant contained the noise determiner /te/.12Two of the six base sentences varied the determiner occurring with a transitive subject, two variedthe determiner occurring with a transitive object, and two were intransitive (therefore varying thedeterminer occurring with the subject). Sentences were presented by audiotape recorder, and theexperimenter recorded responses on a response sheet. Participants had 4 s in which to respond to eachtest item (although participants were allowed a little extra time by pausing the tape player). All testsentences were novel and had not occurred in the exposure set.4.1.5.3. General Grammar Test. The third test examined what participants had learned about aspects ofthe language that were always represented consistently in the input. Specifically, this test examinedwhether participants thought that sentences required verbs, and if they knew that some verbs (thetransitives) required two nouns and others (the intransitives) allowed only one noun. In this task, participantslistened to 16 sentences and were asked to judge each using the same set of faces used in theprevious task. The 16 sentences were actually two versions of each of eight sentences, one grammaticaland the other ungrammatical. For the transitive sentences with missing arguments it was alwaysthe object that was missing. All nouns occurred with determiners. Test sentences were randomizedwith the constraint that the two versions of the same sentence could not follow each other. Randomizationwas the same for all participants. Sentences were played on an audio tape recorder, and theexperimenter recorded the responses on a response sheet. As in the determiner manipulation judgmenttask, participants were given 4 s in which to provide a rating, although the tape was pausedto allow them to respond if needed. Again, all test sentences were novel; none appeared in the exposureset.4.2. ResultsAs above, we present the results from the general grammar test first.4.2.1. General Grammar TestAs in Experiment 1, this test examined participants’ knowledge of parts of the grammar other thandeterminers and served to ensure that learners successfully acquired those parts of the grammar thatwere represented consistently in the input. The test examined participants’ knowledge of sentenceconstruction (did they know that a verb is required in every sentence?) and verb subcategorization(did they know that one set of verbs is transitive, requiring two nouns, and another intransitive,requiring only one noun?).Fig. 9 shows the mean ratings given to grammatical and ungrammatical strings for child and adultparticipants. Two children, one in the 100% condition, one in the 60% + 4ND condition, did not contributedata for this task or the determiner judgment task. Both responded prior to hearing each sentenceand so their data were excluded. We subjected the data to a repeated-measures ANOVA with rule type12 While it would have been desirable to have all four types of test items in all conditions, children were unable to attend to sucha lengthy test. We retained test items with no determiner for all participants, since this would permit comparison with Experiment1 and also would allow us to assess how children responded to having vocabulary training of nouns with no determiners. Acrossconditions, we can assess how children respond to all four types of test items.C.L. Hudson Kam, E.L. Newport / Cognitive Psychology xxx (2009) xxx–xxx 25ARTICLE IN PRESSPlease cite this article in press as: Hudson Kam, C. L., & Newport, E. L. Getting it right by getting it wrong: Whenlearners change languages. Cognitive Psychology (2009), doi:10.1016/j.cogpsych.2009.01.001(two levels) and grammaticality (two levels) as within-subject factors and input type (four levels) andage group (two levels) as between-subjects factors. Grammaticality was significant (F(1,36) = 64.69,p < .0001); participants rated grammatical strings higher than ungrammatical ones. Rule type was alsosignificant (F(1,36) = 18.23, p < .0001), with mean ratings for items testing verb subcategorization higherthan those testing sentence structure. The main effect of age group was significant (F(1,36) = 4.48,p = .041); reflecting the fact that children’s overall mean ratings were lower than adults’. The interactionbetween age and grammaticality was significant (F(1,36) = 7.70, p = .009) because children differentiatedless between grammatical and ungrammatical strings than adults, particularly for thestrings testing basic sentences structure (rule#p#分页标题#e#