On throwing out the baby with the bathwater: A reply to Black and Wilensky's evaluation of story grammars

A number of criticisms of a recent paper byare made. (1) In attempting to assess the observational adequacy of story grammars, they state that a context-free grammar cannot handle discontinuous elements; however, they do not show that such elements occur in the domain to which the grammars apply. Further, they do not present adequate evidence for their claim that there are acceptable stories not accounted for by existing grammars and that the grammars will accept nonstories such as procedures. (2) They state that it has been proven that under natural conditions children cannot learn transformational grammars, which is a misrepresentation of the learnability proofs which have been offered. (3) Most important, they take an unduly narrow approach to story understanding by claiming that people only understand story content and do not have knowledge of story structure which is useful in comprehension or memory. Counterevidence from the literature is cited which indicates that such knowledge is both useful and used, and a number of methods for assessing the psychological adequacy of structural models are discussed.     

A re-evaluation of story grammars

Black and Wilensky (1979) have made serious methodological errors in analyzing story grammars, and in the process they have committed additional errors in applying formal language theory. Our arguments involve clarifying certain aspects of knowledge representation crucial to a proper treatment of story understanding. Particular criticisms focus on the following shortcomings of their presentation: 1) an erroneous statement from formal language theory, 2) misapplication of formal language theory to story grammars, 3) unsubstantiated and doubtful analogies with English grammar, 4) various non sequiturs concerning the generation of non-stories, 5) a false claim based on the artificial distinction between syntax and semantics, and 6) misinterpretation of the role of story grammars in story understanding. We conclude by suggesting appropriate criteria for the evaluation of story grammars.     

Intentional control based on familiarity in artificial grammar learning

It is commonly held that implicit learning is based largely on familiarity. It is also commonly held that familiarity is not affected by intentions. It follows that people should not be able to use familiarity to distinguish strings from two different implicitly learned grammars. In two experiments, subjects were trained on two grammars and then asked to endorse strings from only one of the grammars. Subjects also rated how familiar each string felt and reported whether or not they used familiarity to make their grammaticality judgment. We found subjects could endorse the strings of just one grammar and ignore the strings from the other. Importantly, when subjects said they were using familiarity, the rated familiarity for test strings consistent with their chosen grammar was greater than that for strings from the other grammar. Familiarity, subjectively defined, is sensitive to intentions and can play a key role in strategic control.     

Cognitive structures in comprehension and memory of narrative discourse

The effects of structure and content variables on memory and comprehension of prose passages were studied in two experiments. The experimental passages exemplify a class of simple narrative stories that is described by a generative grammar of plot structures. A comprehension model is proposed that assumes a hierarchical organizational framework of stories in memory, determined by the grammar, representing the abstract structural components of the plot. The quality and characteristics of subjects' memory for stories were tested on a variety of experimental tasks in which story organization was manipulated. Comprehensibility and recall were found to be a function of the amount of inherent plot structure in the story, independent of passage content. Recall probability of individual facts from passages depended on the structural centrality of the facts: Subjects tended to recall facts corresponding to high-level organizational story elements rather than lower-level details. In addition, story summarizations from memory tended to emphasize general structural characteristics rather than specific content. For successively presented stories, both structure and content manipulations influenced recall. Furthermore, repeating story structure across two passages produced facilitation in recall of the second passage, while repeating story content produced proactive interference. The implications for a model of memory for narrative discourse are discussed.     

Ironic effects of drawing attention to story errors

Readers learn errors embedded in fictional stories and use them to answer later general knowledge questions (Marsh, Meade, & Roediger, 2003). Suggestibility is robust and occurs even when story errors contradict well-known facts. The current study evaluated whether suggestibility is linked to participants' inability to judge story content as correct versus incorrect. Specifically, participants read stories containing correct and misleading information about the world; some information was familiar (making error discovery possible), while some was more obscure. To improve participants' monitoring ability, we highlighted (in red font) a subset of story phrases requiring evaluation; readers no longer needed to find factual information. Rather, they simply needed to evaluate its correctness. Readers were more likely to answer questions with story errors if they were highlighted in red font, even if they contradicted well-known facts. Although highlighting to-be-evaluated information freed cognitive resources for monitoring, an ironic effect occurred: Drawing attention to specific errors increased rather than decreased later suggestibility. Failure to monitor for errors, not failure to identify the information requiring evaluation, leads to suggestibility.     

Ironic effects of drawing attention to story errors

Readers learn errors embedded in fictional stories and use them to answer later general knowledge questions (Marsh, Meade, & Roediger, 2003). Suggestibility is robust and occurs even when story errors contradict well-known facts. The current study evaluated whether suggestibility is linked to participants' inability to judge story content as correct versus incorrect. Specifically, participants read stories containing correct and misleading information about the world; some information was familiar (making error discovery possible), while some was more obscure. To improve participants' monitoring ability, we highlighted (in red font) a subset of story phrases requiring evaluation; readers no longer needed to find factual information. Rather, they simply needed to evaluate its correctness. Readers were more likely to answer questions with story errors if they were highlighted in red font, even if they contradicted well-known facts. Although highlighting to-be-evaluated information freed cognitive resources for monitoring, an ironic effect occurred: Drawing attention to specific errors increased rather than decreased later suggestibility. Failure to monitor for errors, not failure to identify the information requiring evaluation, leads to suggestibility.     

Parsing to Learn

Learning a language by parameter setting is almost certainly less onerous than composing a grammar from scratch. But recent computational modeling of how parameters are set has shown that it is not at all the simple mechanical process sometimes imagined. Sentences must be parsed to discover the properties that select between parameter values. But the sentences that drive learning cannot be parsed with the learner's current grammar. And there is not much point in parsing them with just one new grammar. They must apparently be parsed with all possible grammars, in order to find out which one is most successful at licensing the language. The research task is to reconcile this with the fact that the human sentence parsing mechanism, even in adults, has only very limited parallel parsing capacity. I have proposed that all possible grammars can be folded into one, if parameter values are fragments of sentential tree structures that the parser can make use of where necessary to assign a structure to an input sentence. However, the problem of capacity limitations remains. The combined grammar will afford multiple analyses for some sentences, too many to be computed on-line. I propose that the parser computes only one analysis per sentence but can detect ambiguity, and that the learner makes use of unambiguous input only. This provides secure information but relatively little of it, particularly at early stages of learning where few grammars have been excluded and ambiguity is rife. I consider three solutions: improving the parser's ability to extract unambiguous information from partially ambiguous sentences, assuming default parameter values to temporarily eliminate ambiguity, reconfiguring the parameters so that some are subordinate to others and do not present themselves to the learner until the others have been set. A more radical alternative is to give up the quest for error-free learning and permit parameters to be set without regard for whether the parser may have overlooked an alternative analysis of the sentence. If it can be assumed that the human parser keeps a running tally of the parameter values it has accessed, then the learner would do nothing other than parse sentences for comprehension, as adults do. The most useful parameter values would become more and more easily accessed; the noncontributors would drop out of the running. There would be no learning mechanism at all, over and above the parser. But how accurate this system would be remains to be established.     

Employing emotions to drive plot generation in a computer-based storyteller

Emotions are an integral part of the creative process; however, it is hard to find computer models of creativity where emotions play a fundamental role. This paper describes a computer model for plot generation based on emotions and tensions between characters. In particular, the document illustrates how emotions are employed to progress a story in a coherent way and generate novel situations, and how the dramatic tension of the story in progress can be employed to evaluate its interestingness. The model is implemented in a computer program named MEXICA [Pérez y Pérez, R., & Sharples, M. (2001). MEXICA: a computer model of a cognitive account of creative writing. Journal of Experimental and Theoretical Artificial Intelligence, 13(2), 119–139]; this work concentrates on the role of emotions in plot generation. The main claim is that a story can be represented as a cluster or group of emotional links and tensions between characters that progresses over story-time; story-actions work as operators that modify such clusters. I present results showing how story generation is affected by various model parameters. This approach means the program is flexible, as it avoids using predefined story-structures or characters’ goals to drive story generation. Furthermore, evaluation of computer generated stories showed that MEXICA’s stories were most often selected as the best story. This suggests that the story-generation mechanisms within MEXICA are sufficiently rich to generate interesting and novel stories.     

Perceptual constraints and the learnability of simple grammars

Cognitive processes are often attributed to statistical or symbolic general-purpose mechanisms. Here we show that some spontaneous generalizations are driven by specialized, highly constrained symbolic operations. We explore how two types of artificial grammars are acquired, one based on repetitions and the other on characteristic relations between tones ("ordinal" grammars). Whereas participants readily acquire repetition-based grammars, displaying early electrophysiological responses to grammar violations, they perform poorly with ordinal grammars, displaying no such electrophysiological responses. This outcome is problematic for both general symbolic and statistical models, which predict that both types of grammars should be processed equally easily. This suggests that some simple grammars are acquired using perceptual primitives rather than general-purpose mechanisms; such primitives may be elements of a "toolbox" of specialized computational heuristics, which may ultimately allow constructing a psychological theory of symbol manipulation.     

Telling a “baby story”: Mothers narrating their pre-schoolers' past across two cultural contexts

Mothers from two middle-class contexts from Berlin, Germany (n = 35), and Delhi, India (n = 28) told a baby story to their 3-year olds about an event that had happened during the children's first year of life. The contexts represented two cultural models: the model of psychological autonomy (Berlin) and the model of autonomy-relatedness (Delhi). We investigated the culture-specific functions of this reminiscing task as reflected in the structure, content and specificity of the stories. The stories in both contexts were minimally interactive and the children contributed few elaborations themselves. Stories told by the Berlin mothers were longer, and more specific. Mothers in both contexts were similarly elaborative relative to being repetitive. The stories were highly child-centred in both contexts but even more child-centred for Delhi. Importantly, maternal narrations from the Berlin context were embedded in a frame story that characterised the child's individual past; stories thus constructed “exclusive baby stories”. Most stories told by the Delhi mothers had no frame story but instead were about what the child used to do as a baby in general; they thus constructed “routine baby stories”. Results are interpreted in the view of the underlying self, social and directive functions of this reminiscing task.     

Causal Attribution and Perceived Realism of Stories

Just as people generate causal explanations for social events around them, story readers usually generate inferences about causality of events when reading a story. The attribution literature suggests that, when judging events that happen to others, people spontaneously generate dispositional explanations for negative events and situational explanations for positive events and the reverse when judging events that happen to themselves. Three experiments examined how these spontaneously generated inferences of causality interacted with causal explanations provided by the text of a story to influence perceived realism. The results indicate that the relationship between spontaneously generated causal attributions and information supplied by the story had little influence on realism judgments about story characters or about other people. When evaluating the story scenario for the self, however, results of all three experiments show that people find information consistent with their own spontaneous attributions more realistic. The results contribute to our understanding of the psychological processes that may drive realism evaluations of stories and possible contrasting mechanisms between attributions in story worlds and the social world.     

Empirical tests of aspects of a theory of love as a story

We provide empirical tests of aspects of a theory of love as a story. According to this theory, people develop—as an interaction between their personality and their experiences—stories of what they believe loving relationships should be. Examples of such stories are addiction, mystery, police, and travel stories. They then seek out and find greatest satisfaction with partners whose stories correspond more closely with their own. The data from two studies indicate that the theory and instrument have some promise for understanding people's ways of conceptualizing love. In particular, couples involved in intimate relationships tend to share similar profiles of love stories; the more similar the stories of two members of a couple, the more likely they are to be satisfied with their relationship. Copyright © 2001 John Wiley & Sons, Ltd.     

The nature of the memory buffer in implicit learning: Learning Chinese tonal symmetries

Previous research has established that people can implicitly learn chunks, which (in terms of formal language theory) do not require a memory buffer to process. The present study explores the implicit learning of nonlocal dependencies generated by higher than finite-state grammars, specifically, Chinese tonal retrogrades (i.e. centre embeddings generated from a context-free grammar) and inversions (i.e. cross-serial dependencies generated from a mildly context-sensitive grammar), which do require buffers (for example, last in-first out and first in-first out, respectively). People were asked to listen to and memorize artificial poetry instantiating one of the two grammars; after this training phase, people were informed of the existence of rules and asked to classify new poems, while providing attributions of the basis of their judgments. People acquired unconscious structural knowledge of both tonal retrogrades and inversions. Moreover, inversions were implicitly learnt more easily than retrogrades constraining the nature of the memory buffer in computational models of implicit learning.     

Not evidence for separable controlled and automatic influences in artificial grammar learning: comment on Higham, Vokey, and Pritchard (2000)

P. A. Higham, J. R. Vokey, and J. L. Pritchard (2000) claimed to provide evidence for separable controlled and automatic processes in artificial grammar learning. It is argued that their results are compatible with a single controlled influence: Participants might mistakenly identify more grammatical items than nongrammatical items as belonging to the other grammar, because the grammars are very similar to each other, and the nongrammatical items are relatively highly dissimilar. Participants' knowledge may be ambiguous, rather than automatic. It is further argued that even if Higham et al.'s data do support automatic effects, opposition logic, in this case, cannot be said to have succeeded where dissociation logic has failed, because it is used to address the issue of whether participants have conscious control over the knowledge they acquire, rather than whether they possess conscious awareness of that knowledge.     

Poetry as a Mnemonic Prompt in Children's Stories

This research investigates the effects of prosody on children's recall for stories using two successive studies. Study 1 is an ethnographic exploration of a group of fifth graders creating summaries of a children's story with overt prosodic elements. Study 2 is a quasi-experiment in which the story summaries created by students who heard one of two versions (more or less prosodic) of the story from Study 1 were compared. Overall, we find that the amount of prosody in a story has a significant effect on children's story recall and the quality of the stories they produce.     

Story processing in right-hemisphere brain-damaged patients.

The understanding of stories requires sensitivity to structural aspects of narrative, the emotional content conveyed by the narrative, and the interaction between structural and emotional facets of the story. Right-hemisphere-damaged (RHD) and normal control subjects performed a number of different analytic tasks which probed their competence at story comprehension. Results revealed that RHD subjects perform at a level comparable to that of normal controls with stories that follow a canonical form and that they show few difficulties with structural aspects of narrative. Contrary to expectation, they are strongly influenced by the "interest" level of a story and by other factors that tap emotional sensitivity. Findings are discussed in terms of the processing and arousal mechanisms which may give rise to the observed pattern of difficulties in RHD patients.     

A Universe of Stories

The complete adoption of a narrative paradigm for therapy is proposed and its implications and possibilities spelled out, especially for life in a post-modern world that lacks any objective frame of reference. In the post-modern sensibility, the story is set free to perform as simply a story that allows for re-invention as the story-teller finds a voice rooted in the person's own experience and in the connection of her story to those of others, and to larger stories of culture and humanity. The realization that we are all characters in each other's stories as well as our own reminds us that our stories only go forward as we act in ways that also forward the stories of others. A hermeneutics for such therapy validates and questions all points of view so that they do not harden into stories that are assumed to be objectively true.     

Whose idea is it anyway? The importance of reputation in acknowledgement

Six‐year‐old children negatively evaluate plagiarizers just as adults do (Olson & Shaw, 2011), but why do they dislike plagiarizers? Children may think plagiarism is wrong because plagiarizing negatively impacts other people's reputations. We investigated this possibility by having 6‐ to 9‐year‐old children evaluate people who shared their own or other people's ideas (stories). In Experiment 1, we found that children consider it acceptable to retell someone else's story if the source is given credit for their story (improving the source's reputation), but not if the reteller claims credit for the story (steals credit away from someone else). Experiments 2 and 3 showed that children do not consider it bad to lie by giving someone else credit for one's own good story (improving someone else's reputation), but do consider it bad to give someone else credit for one's own bad story (improving one's own reputation at the expense of someone else's). Experiment 4 demonstrated that children think it is equally bad to take credit for someone else's idea for oneself as it is to take someone else's idea and give credit to someone else, suggesting that children dislike when others take credit away from someone else, regardless of whether or not it improves the plagiarizer's reputation. Our results suggest that children dislike plagiarism because it negatively affects others' reputations by taking credit away from them.     

Individuality of Narratives Generated by Individuals: An Investigation Using Four-Panel Comics

In this study, we conducted an empirical research focusing on the generation of stories to obtain a hypothesis on how different the stories that people generate in their daily lives by linking events and situations are. For this purpose, five illustrations that revealed only the first and fourth panels of a four-panel comic were used as stimuli to generate a narrative in which the participants could make a connection between them. A total of 40 stories were obtained from eight participants. As a result, the narrative emerged in a variety of ways, even in events with little personal connection to the participants. This reveals that there are individual differences in the way people generate narratives themselves, despite their relationship with and position on the event. When the stories were typified, none of the participants generated similar stories in all five illustrations. This suggests that, because the combination of stories a person creates about every event is unique, there is only one story about the world, and one about himself/herself.