Transfer in artificial grammar learning: The role of repetition information.

In this article, the authors report 2 experiments that investigated the sources of information used in transfer and nontransfer tasks in artificial grammar learning. Multiple regression analyses indicated that 2 types of information about repeating elements were crucial for performance in both tasks: information about the repetition of adjacent elements and information about repetition of elements in the whole item. Similarity of test items to specific training items and chunk information influenced participants' judgments only in nontransfer tasks. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Theories of artificial grammar learning.

Artificial grammar learning (AGL) is one of the most commonly used paradigms for the study of implicit learning and the contrast between rules, similarity, and associative learning. Despite five decades of extensive research, however, a satisfactory theoretical consensus has not been forthcoming. Theoretical accounts of AGL are reviewed, together with relevant human experimental and neuroscience data. The author concludes that satisfactory understanding of AGL requires (a) an understanding of implicit knowledge as knowledge that is not consciously activated at the time of a cognitive operation; this could be because the corresponding representations are impoverished or they cannot be concurrently supported in working memory with other representations or operations, and (b) adopting a frequency-independent view of rule knowledge and contrasting rule knowledge with specific similarity and associative learning (co-occurrence) knowledge. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Associative chunk strength in artificial grammar learning.

In 4 experiments, adherence to grammatical rules and associative chunk strength (including different measures, each calculated on the basis of the frequency with which bigrams and trigrams present in the test strings appeared in the learning strings) were manipulated independently in the test phase of an artificial grammar learning task. When participants learned few items of the grammar (Experiments 1A and 2A), the associated items were more often classified as grammatical than the nonassociated ones. On the other hand, when the learning phase included most of the grammatical items (Experiments 1B and 2B), the only effect observed was an effect of grammaticality. These results suggest that, depending on the specific constraints of the tasks, knowledge based on bigrams and trigrams and knowledge based on the abstraction of the grammatical structure can be used for the classification task. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The role of similarity in artificial grammar learning.

The authors examine the role of similarity in artificial grammar learning (AGL; A. S. Reber, 1989). A standard finite-state language was used to create stimuli that were arrangements of embedded geometric shapes (Experiment 1), connected lines (Experiment 2), and sequences of shapes (Experiment 3). Main effects for well-known predictors from the literature (grammaticality, associative global and anchor chunk strength, novel global and anchor chunk strength, length of items, and edit distance) were observed, thus replicating previous work. However, the authors extend previous research by using a widely known similarity-based exemplar model of categorization (the generalized context model; R. M. Nosofsky, 1989) to fit grammaticality judgments, by nested regression analyses. The results suggest that any explanation of AGL that is based on the existing theories is incomplete without a similarity process as well. Also, the results provide a foundation for further interpreting AGL in the wider context of categorization research. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Two modes of transfer in artificial grammar learning.

Participants can transfer grammatical knowledge acquired implicitly in 1 vocabulary to new sequences instantiated in both the same and a novel vocabulary. Two principal theories have been advanced to account for these effects. One suggests that sequential dependencies form the basis for cross-domain transfer (e.g., Z. Dienes et al, see record 1999-13687-003). Another argues that a form of episodic memory known as abstract analogy is sufficient (e.g., L. R. Brooks & J. R. Vokey, see record 1992-00385-001). Three experiments reveal the contributions of the 2. In Experiment 1 sequential dependencies form the only basis for transfer. Experiment 2 demonstrates that this process is impaired by a change in the distributional properties of the language. Experiment 3 demonstrates that abstract analogy of repetition structure is relatively immune to such a change. These findings inform theories of artificial grammar learning and the transfer of grammatical knowledge. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The information acquired during artificial grammar learning.

In an artificial grammar learning task, amnesic patients classified test items as well as normal Ss did. Item similarity did not affect grammaticality judgments when similar and nonsimilar test items were balanced for the frequency with which bigrams and trigrams (chunks) that appeared in the training set also appeared in the test items. Amnesic Ss performed like normal Ss. Results suggest that concrete information about letter chunks can influence grammaticality judgments and that this information is acquired implicitly. The similarity of whole test items to training items does not appear to affect grammaticality judgments. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Dissociations of grammaticality and specific similarity effects in artificial grammar learning.

Three artificial grammar learning experiments investigated the memory processes underlying classification judgments. In Experiment 1, effects of grammaticality, specific item similarity, and chunk frequency were analogous between classification and recognition tasks. In Experiments 2A and 2B, instructions to exclude "old" and "similar" test items, under conditions that limited the role of conscious recollection, dissociated grammaticality and similarity effects in classification. Dividing attention at test also produced a dissociation in Experiment 3. It is concluded that a dual-process model of classification, whereby the grammaticality and specific similarity effects are based mostly on automatic and intentional memory processes, respectively, is consistent with the results, whereas a unitary mechanism account is not. This conclusion is further supported by evidence indicating that chunk frequency had both implicit and explicit influences on classification judgments. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Beyond dissociation logic: Evidence for controlled and automatic influences in artificial grammar learning.

Evidence for unconscious learning has typically been based on dissociations between direct and indirect tests of learning. Because of some inherent problems with dissociation logic, we applied the logic of opposition to 2 artificial grammar learning experiments. In Experiment 1, participants were exposed to 2 different sets of letter strings, generated from 2 different grammars, and later rated test strings for grammaticality with either in-concert (rate grammatical strings consistent with either structure) or opposition (rate grammatical only strings from 1 of the structures) instructions. Manipulating response deadline affected controlled, but not automatic influences. In Experiment 2, after similar training, a source-monitoring test was administered from which the in-concert and opposition conditions were derived. The test indicated that varying the retention interval affected controlled, but not automatic, influences. The results are discussed in terms of awareness, knowledge representation, and metacognitive processing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Independent judgment-linked and motor-linked forms of artificial grammar learning

Actinosporean infection of the oligochaete fauna living in the mud and on the vegetation of fish ponds used for rearing common carp in polyculture was studied during a one-year survey at a fish farm in Hungary, located south of Budapest. Twenty-eight actinospore types were isolated from the oligochaetes Tubifex tubifex, Branchiura a sowerbyi, Limnodrilus hoffmeisteri, Nais elinguis, and Stylaria lacustris collected during the survey, which could be classified into the triactinomyxon, raabeia, aurantiactinomyxon and neoactinomyxon groups. Drawings depicting individual actinospore types are presented on plates and their characteristic dimensions have been summarised in tables. The prevalence and seasonality of actinosporean infections observed in the various oligochaete species have been illustrated graphically. Infection by actinospores showed a pronounced seasonality. In the spring, summer and autumn the prevalence of raabeia infection in Branchiura exceeded 90%, while in the winter it dropped to 42%. A similar phenomenon could be observed for aurantiactinomyxon infection, while neoactinomyxon infection reached its peak in the autumn. In Tubifex, Limnodrilus, Nais and Stylaria species the peak of actinosporean infection occurred, with minor differences, in the spring and summer. Actinosporean infection in the individual Oligochaeta species showed much higher prevalence values than had been reported in the literature, which can be explained by the novelty of the examination technique used. It cannot be decided with absolute certainty which myxosporean developmental stage the different actinospore types described during this survey represent of the species of Myxosporea parasitic in the given ponds. This would require an experimental study for which the data presented here may serve as a basis.     

Chunks are not enough: The insufficiency of feature frequency-based explanations of artificial grammar learning.

Two experiments with a total of 104 undergraduates tested chunk frequency explanations of artificial grammar learning, which hold that classification performance is dependent on some metric derived from the frequency with which certain features occur within the letter string stimuli. Exp 1 revealed that classification performance was affected by close graphemic similarity between specific training (e.g., MXRVXT) and test strings (e.g., MXRMXT), despite the fact that similar strings did not contain frequently occurring features (e.g., bigrams or trigrams). This effect was replicated in Exp 2a, and Exp 2b demonstrated that substituting letters to make the consonant strings pronounceable (e.g., substituting X, R, and T, in the consonant string MXRMXT with Y, A, I, to produce MYAMYI) affected classification performance, despite the fact that objective measures of feature frequency were not altered. It is argued that models of classification that focus entirely on the frequency of features within the literal stimulus are insufficient, and that some allowance must be made for how the stimulus is encoded. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The transfer effect in artificial grammar learning: Reappraising the evidence on the transfer of sequential dependencies.

Exposure to sequences of elements generated by an artificial grammar enables observers to classify new sequences presented either with the same (within-domain) or different (across-domain) vocabulary elements as being well- or ill-formed according to that grammar (A. S. Reber, 1969). Experiment 1 replicated G. T. M. Altmann, Z. Dienes, and A. Goode's (1995) demonstration of this effect, but inspection of hits and correct rejections revealed that a single cue was used to reject a subset of the ungrammatical sequences. Experiment 2 removed this cue, and participants no longer discriminated between grammatical and ungrammatical sequences in the novel domain. The authors conclude that G. T. M. Altmann et al.'s demonstration of discrimination in the novel domain did not necessitate the application of knowledge pertaining to sequential dependencies. Implications of the data for other studies, showing abstraction of knowledge across test sequences, are also considered. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Transposition: A reevaluation.

Theories of transposition are evaluated in terms of the available empirical data. The most often cited model, that of Spence (1937), is found to be adequate in its explanation of simple transposition and the distance effect. It encounters difficulty in explaining or incorporating the findings of: intermediate size, simultaneous versus successive discrimination, contrast and background effects, and mediated transposition. Other models are considered, with adaptation-level approaches providing the most adequate alternative to the Spence position. (3 p. ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Abstractness of implicit grammar knowledge: Comments on Perruchet and Pacteau's analysis of synthetic grammar learning.

On the basis of 3 experiments, P. Perruchet and C. Pacteau (see record 1991-00329-001) argued that implicitly acquired knowledge of a synthetic grammar consists of little more than knowledge of pairwise associations between pairs of letters in the grammar. By comparing their results with a study by R. C. Mathews et al (see record 1990-03549-001), it is argued that (a) implicitly acquired knowledge is much richer and more abstract than suggested by Perruchet and Pacteau, (b) their recognition measures are less sensitive than the recall measures of Mathews et al for detecting conscious awareness of implicit knowledge, and (c) fragmentary knowledge of a grammar constitutes abstract rules that enable performance of complex tasks when integrated into a system for combining knowledge across rules. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Base-rate neglect in ALCOVE: A critical reevaluation.

A recent hybrid model of categorization (Attention Learning Covering Map [ALCOVE]; J. K. Kruschke, 1992) has combined the most desirable properties of exemplar models with a connectionist architecture and learning rule. A critically important property of ALCOVE is its apparent ability to account for base-rate neglect, a phenomenon beyond the purview of previous exemplar models. This article reexamines ALCOVE's base-rate neglect predictions and shows that they are confined to a very limited set of circumstances. In most cases, ALCOVE is unable to produce base-rate neglect. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Indirect effects of synthetic grammar learning in an identification task.

Two experiments investigated the effects of incidental learning of an artificial grammar on an indirect measure. Ss memorized consonant strings and later identified these in a perceptual clarification procedure. The competitive chunking model by E. Servan-Schreiber and J. R. Anderson (see record 1990-27513-001) predicted faster identifications of grammatical as opposed to nongrammatical strings. Both experiments confirmed this prediction. Exp 2 further investigated whether faster identification induces a feeling of familiarity, which should increase the probability of subjects responding "grammatical" to a string. Faster identifications were indeed related to "old" responses in a recognition judgment. However, there was no systematic relation between speed of identification and grammaticality judgments, which is inconsistent with the prediction of the competitive chunking model that familiarity exclusively mediates grammaticality judgments. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Transfer learning in chronic schizophrenics.

Compared reversal shift, where the assignment of stimuli to categories changed but the relevant dimension did not change, and extradimensional shift, where a different dimension was made relevant after training. 28 chronic schizophrenics were trained to the same consecutive correct response criterion in a 3-choice form or color discrimination and shifted to 1 of 4 possible transfer tasks. It was found that (a) extradimensional shift took more trials and errors to criterion than reversal shift, and (b) more perseveration as compared to other errors were made by the total sample. The need for a uniform methodology and an evaluation of S selection factors when investigating the performance of schizophrenics is discussed. (23 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Extrinsic incentives and token-based programs: A reevaluation.

Criticizes F. M. Levine and G. Fasnacht's (see record 1975-07966-001) recommendation against the use of token programs (TPs) and extrinsic incentives in attribution theory and generalization research. Most TPs cited by Levine and Fasnacht successfully produced behavior change, had few observable adverse effects on Ss or their behavior, and required substantial improvement only in their procedures for programming generalization. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Cognitive desensitization and avoidance behavior: A reevaluation.

Attempted to partially replicate and extend a study by S. Valins and A. Ray (see record 1968-04107-001) on the effects of false heart rate feedback on avoidance behavior. In the previous study, Ss reporting a fear of snakes saw slides of snakes while hearing either meaningless sounds or bogus heart rate feedback indicating they were not afraid of snakes. The present study with 30 female undergraduates replicated these conditions while including a no-treatment control and a group of Ss led to believe that snakes did affect them. Actual heart rate was monitored during the sessions and posttreatment interviews were conducted to more fully assess Ss' reactions. Analyses failed to reveal significant between-group differences on any of the outcome indices of fear. This was attributed to the apparent failure of experimental Ss to reevaluate their fear on the basis of their heart rate feedback. (15 ref.) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Risk taking and vocational choice: A reevaluation.

Tested whether Ss preferring high levels of risk tend to choose high-risk occupations more often than low-risk occupations when asked to choose between them. Three separate risk-taking measures were used to divide 96 college freshmen and sophomores into high and low risk takers. Degree of occupational risk was defined by having Ss rate occupational titles according to perceived vocational success and probable income. Three Type III analyses of variance, one for each risk-taking measure, were carried out on the number of high-risk occupations chosen. Results show that risk-taking preferences were not significantly related to Ss' vocational choices on the selection tasks. Subsequent analyses compared the risk preferences of vocationally decided and undecided Ss, but no significant differences were found. The relationship of findings to previous research is discussed. (15 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)