RESUMO
Whereas principles of communicative efficiency and legal doctrine dictate that laws be comprehensible to the common world, empirical evidence suggests legal documents are largely incomprehensible to lawyers and laypeople alike. Here, a corpus analysis (n = 59) million words) first replicated and extended prior work revealing laws to contain strikingly higher rates of complex syntactic structures relative to six baseline genres of English. Next, two preregistered text generation experiments (n = 286) tested two leading hypotheses regarding how these complex structures enter into legal documents in the first place. In line with the magic spell hypothesis, we found people tasked with writing official laws wrote in a more convoluted manner than when tasked with writing unofficial legal texts of equivalent conceptual complexity. Contrary to the copy-and-edit hypothesis, we did not find evidence that people editing a legal document wrote in a more convoluted manner than when writing the same document from scratch. From a cognitive perspective, these results suggest law to be a rare exception to the general tendency in human language toward communicative efficiency. In particular, these findings indicate law's complexity to be derived from its performativity, whereby low-frequency structures may be inserted to signal law's authoritative, world-state-altering nature, at the cost of increased processing demands on readers. From a law and policy perspective, these results suggest that the tension between the ubiquity and impenetrability of the law is not an inherent one, and that laws can be simplified without a loss or distortion of communicative content.
Assuntos
Idioma , Humanos , Feminino , Masculino , Redação , Adulto , Comunicação , CompreensãoRESUMO
Accounts of human language comprehension propose different mathematical relationships between the contextual probability of a word and how difficult it is to process, including linear, logarithmic, and super-logarithmic ones. However, the empirical evidence favoring any of these over the others is mixed, appearing to vary depending on the index of processing difficulty used and the approach taken to calculate contextual probability. To help disentangle these results, we focus on the mathematical relationship between corpus-derived contextual probability and the N400, a neural index of processing difficulty. Specifically, we use 37 contemporary transformer language models to calculate the contextual probability of stimuli from 6 experimental studies of the N400, and test whether N400 amplitude is best predicted by a linear, logarithmic, super-logarithmic, or sub-logarithmic transformation of the probabilities calculated using these language models, as well as combinations of these transformed metrics. We replicate the finding that on some datasets, a combination of linearly and logarithmically-transformed probability can predict N400 amplitude better than either metric alone. In addition, we find that overall, the best single predictor of N400 amplitude is sub-logarithmically-transformed probability, which for almost all language models and datasets explains all the variance in N400 amplitude otherwise explained by the linear and logarithmic transformations. This is a novel finding that is not predicted by any current theoretical accounts, and thus one that we argue is likely to play an important role in increasing our understanding of how the statistical regularities of language impact language comprehension.
RESUMO
This paper presents a collaboration between a clinician (C.M.J.) and a research team (W.B., B.M., and S.M.) to address the question: At an operational level, what happens in the special form of conversation that is psychotherapy? How can we study, beyond a priori lenses of psychoanalytic models, what we are actually doing when we engage in this process? How can we capture from the linear flow of conversation, the simultaneous, complex, active, interwoven, dimensional emotion schemas that words can only point toward? To address the question, we first present the need for new approaches in the current climate within the clinical and research communities. Next, we address the challenges for clinicians and researchers by using multiple code theory and derived linguistic measures that offer an objective view of the processes of subjectivity. We then apply the research methods to the clinical data to illustrate the yield of the collaborative effort-a yield that captures the connection between the linear flow of words and the arousal, verbal expression, and reflection/integration of emotion schemas without the usual filters of psychoanalytic models of process and change. The project illustrates the critical value of clinicians' perspectives to guide researchers and encourages clinicians to participate in research to advance our field. For researchers, this project represents a "fourth generation" of process research that includes the criteria of video-recorded, transcribed data; the clinician's report of their experience; a theory of how emotion-laden meaning and motivations (emotion schemas) are expressed in the therapeutic conversation; and reliable, valid measures to capture and represent those processes; and that encourages researchers to access the rich contributions of clinicians' understanding. The implication for clinical practice is a new way to look beyond the lens of psychoanalytic models into what is actually unfolding in real time.
RESUMO
Languages are governed by syntactic constraints-structural rules that determine which sentences are grammatical in the language. In English, one such constraint is subject-verb agreement, which dictates that the number of a verb must match the number of its corresponding subject: "the dogs run", but "the dog runs". While this constraint appears to be simple, in practice speakers make agreement errors, particularly when a noun phrase near the verb differs in number from the subject (for example, a speaker might produce the ungrammatical sentence "the key to the cabinets are rusty"). This phenomenon, referred to as agreement attraction, is sensitive to a wide range of properties of the sentence; no single existing model is able to generate predictions for the wide variety of materials studied in the human experimental literature. We explore the viability of neural network language models-broad-coverage systems trained to predict the next word in a corpus-as a framework for addressing this limitation. We analyze the agreement errors made by Long Short-Term Memory (LSTM) networks and compare them to those of humans. The models successfully simulate certain results, such as the so-called number asymmetry and the difference between attraction strength in grammatical and ungrammatical sentences, but failed to simulate others, such as the effect of syntactic distance or notional (conceptual) number. We further evaluate networks trained with explicit syntactic supervision, and find that this form of supervision does not always lead to more human-like syntactic behavior. Finally, we show that the corpus used to train a network significantly affects the pattern of agreement errors produced by the network, and discuss the strengths and limitations of neural networks as a tool for understanding human syntactic processing.
RESUMO
How listeners weight a wide variety of information to interpret ambiguities in the speech signal is a question of interest in speech perception, particularly when understanding how listeners process speech in the context of phrases or sentences. Dominant views of cue use for language comprehension posit that listeners integrate multiple sources of information to interpret ambiguities in the speech signal. Here, we study how semantic context, sentence rate, and vowel length all influence identification of word-final stops. We find that while at the group level all sources of information appear to influence how listeners interpret ambiguities in speech, at the level of the individual listener, we observe systematic differences in cue reliance, such that some individual listeners favor certain cues (e.g., speech rate and vowel length) to the exclusion of others (e.g., semantic context). While listeners exhibit a range of cue preferences, across participants we find a negative relationship between individuals' weighting of semantic and acoustic-phonetic (sentence rate, vowel length) cues. Additionally, we find that these weightings are stable within individuals over a period of 1 month. Taken as a whole, these findings suggest that theories of cue integration and speech processing may fail to capture the rich individual differences that exist between listeners, which could arise due to mechanistic differences between individuals in speech perception.
RESUMO
One important goal of cognitive science is to understand the mind in terms of its representational and computational capacities, where computational modeling plays an essential role in providing theoretical explanations and predictions of human behavior and mental phenomena. In my research, I have been using computational modeling, together with behavioral experiments and cognitive neuroscience methods, to investigate the information processing mechanisms underlying learning and visual cognition in terms of perceptual representation and attention strategy. In perceptual representation, I have used neural network models to understand how the split architecture in the human visual system influences visual cognition, and to examine perceptual representation development as the results of expertise. In attention strategy, I have developed the Eye Movement analysis with Hidden Markov Models method for quantifying eye movement pattern and consistency using both spatial and temporal information, which has led to novel findings across disciplines not discoverable using traditional methods. By integrating it with deep neural networks (DNN), I have developed DNN+HMM to account for eye movement strategy learning in human visual cognition. The understanding of the human mind through computational modeling also facilitates research on artificial intelligence's (AI) comparability with human cognition, which can in turn help explainable AI systems infer humans' belief on AI's operations and provide human-centered explanations to enhance human-AI interaction and mutual understanding. Together, these demonstrate the essential role of computational modeling methods in providing theoretical accounts of the human mind as well as its interaction with its environment and AI systems.
Assuntos
Cognição , Redes Neurais de Computação , Humanos , Cognição/fisiologia , Movimentos Oculares/fisiologia , Simulação por Computador , Percepção Visual/fisiologia , Atenção/fisiologia , Inteligência Artificial , Aprendizagem/fisiologiaRESUMO
Working memory (WM) capacity has been shown to influence how readers resolve syntactic ambiguities. Building on the work of Swets et al. (2007, Journal of Experimental Psychology: General, 136[1], 64-81), the goal of the present study was to assess the effects of working memory and language proficiency on first language (L1) relative clause attachment decisions across three different language samples: English monolinguals, L1-L2 Spanish-English heritage bilinguals, and L1-L2 Mandarin-English bilinguals. Binomial logistic regression analyses demonstrated that low WM span is associated with a preference to attach ambiguous relative clauses higher in the syntactic structure, as reported by Swets et al. (2007), and contrary to a recency strategy. We also observed that proficiency in L1 and L2 have little effect, suggesting that relative clause attachment preferences primarily reflect the properties of the language and the working memory capacity of the comprehender.
RESUMO
In real-time sentence comprehension, the comprehender is often required to establish syntactic dependencies between words that are linearly distant. Major models of sentence comprehension assume that longer dependencies are more difficult to process because of working memory limitations. While the expected effect of distance on reading times (locality effect) has been robustly observed in certain constructions, such as relative clauses in English, its generalizability to a wider range of constructions has been empirically questioned. The current study proposes a new metric of syntactic distance that capitalizes on the flexible constituency of Combinatory Categorial Grammar (CCG), and argues that it offers a unified account of the locality effects. It is shown that this metric correctly predicts both the presence of the locality effect in English relative clauses and its absence in verb-final languages, without assuming language- or dependency-specific differences in the sensitivity to the locality effect. It is further shown that the CCG-based distance is a significant predictor of the self-paced reading times from an English corpus, even when other known predictors such as dependency-based locality and surprisal are taken into account. These results suggest that human sentence comprehension involves rapid integration of input words into efficiently compressed syntactic representations, and CCG is a plausible theory of the grammar that subserves this process.
RESUMO
In psychophysics and psychometrics, an integral method to the discipline involves charting how a person's response pattern changes according to a continuum of stimuli. For instance, in hearing science, Visual Analog Scaling tasks are experiments in which listeners hear sounds across a speech continuum and give a numeric rating between 0 and 100 conveying whether the sound they heard was more like word "a" or more like word "b" (i.e. each participant is giving a continuous categorization response). By taking all the continuous categorization responses across the speech continuum, a parametric curve model can be fit to the data and used to analyze any individual's response pattern by speech continuum. Standard statistical modeling techniques are not able to accommodate all of the specific requirements needed to analyze these data. Thus, Bayesian hierarchical modeling techniques are employed to accommodate group-level non-linear curves, individual-specific non-linear curves, continuum-level random effects, and a subject-specific variance that is predicted by other model parameters. In this paper, a Bayesian hierarchical model is constructed to model the data from a Visual Analog Scaling task study of mono-lingual and bi-lingual participants. Any nonlinear curve function could be used and we demonstrate the technique using the 4-parameter logistic function. Overall, the model was found to fit particularly well to the data from the study and results suggested that the magnitude of the slope was what most defined the differences in response patterns between continua.
Assuntos
Teorema de Bayes , Modelos Estatísticos , Humanos , Escala Visual Analógica , Psicometria , Psicofísica , Feminino , Percepção da Fala , MasculinoRESUMO
Prior research suggests that the development of speech perception and word recognition stabilises in early childhood. However, recent work suggests that development of these processes continues throughout adolescence. This study aimed to investigate whether these developmental changes are based solely within the lexical system or are due to domain general changes, and to extend this investigation to lexical-semantic processing. We used two Visual World Paradigm tasks: one to examine phonological and semantic processing, one to capture non-linguistic domain-general skills. We tested 43 seven- to nine-year-olds, 42 ten- to thirteen-year-olds, and 30 sixteen- to seventeen-year-olds. Older children were quicker to fixate the target word and exhibited earlier onset and offset of fixations to both semantic and phonological competitors. Visual/cognitive skills explained significant, but not all, variance in the development of these effects. Developmental changes in semantic activation were largely attributable to changes in upstream phonological processing. These results suggest that the concurrent development of linguistic processes and broader visual/cognitive skills lead to developmental changes in real-time phonological competition, while semantic activation is more stable across these ages.
RESUMO
Many studies of human language processing have shown that readers slow down at less frequent or less predictable words, but there is debate about whether frequency and predictability effects reflect separable cognitive phenomena: are cognitive operations that retrieve words from the mental lexicon based on sensory cues distinct from those that predict upcoming words based on context? Previous evidence for a frequency-predictability dissociation is mostly based on small samples (both for estimating predictability and frequency and for testing their effects on human behavior), artificial materials (e.g., isolated constructed sentences), and implausible modeling assumptions (discrete-time dynamics, linearity, additivity, constant variance, and invariance over time), which raises the question: do frequency and predictability dissociate in ordinary language comprehension, such as story reading? This study leverages recent progress in open data and computational modeling to address this question at scale. A large collection of naturalistic reading data (six datasets, >2.2 M datapoints) is analyzed using nonlinear continuous-time regression, and frequency and predictability are estimated using statistical language models trained on more data than is currently typical in psycholinguistics. Despite the use of naturalistic data, strong predictability estimates, and flexible regression models, results converge with earlier experimental studies in supporting dissociable and additive frequency and predictability effects.
RESUMO
Computational models of infant word-finding typically operate over transcriptions of infant-directed speech corpora. It is now possible to test models of word segmentation on speech materials, rather than transcriptions of speech. We propose that such modeling efforts be conducted over the speech of the experimental stimuli used in studies measuring infants' capacity for learning from spoken sentences. Correspondence with infant outcomes in such experiments is an appropriate benchmark for models of infants. We demonstrate such an analysis by applying the DP-Parser model of Algayres and colleagues to auditory stimuli used in infant psycholinguistic experiments by Pelucchi and colleagues. The DP-Parser model takes speech as input, and creates multiple overlapping embeddings from each utterance. Prospective words are identified as clusters of similar embedded segments. This allows segmentation of each utterance into possible words, using a dynamic programming method that maximizes the frequency of constituent segments. We show that DP-Parse mimics American English learners' performance in extracting words from Italian sentences, favoring the segmentation of words with high syllabic transitional probability. This kind of computational analysis over actual stimuli from infant experiments may be helpful in tuning future models to match human performance.
Assuntos
Percepção da Fala , Lactente , Humanos , Estudos Prospectivos , Idioma , Psicolinguística , Fala , Desenvolvimento da Linguagem , Simulação por ComputadorRESUMO
Theories of word processing propose that readers are sensitive to statistical co-occurrences between spelling and meaning. Orthographic-semantic consistency (OSC) measures provide a continuous estimate of the statistical regularities between spelling and meaning. Here we examined Malay, an Austronesian language that is agglutinative. In Malay, stems are often repeated in other words that share a related meaning (e.g., sunyi/quiet; ke-sunyi-an/silence; makan/eat; makan-an/foods). The first goal was to expand an existing large Malay database by computing OSC estimates for 2,287 monomorphemic words. The second goal was to explore the impact of root family size and OSC on lexical decision latencies for monomorphemic words. Decision latencies were collected for 1,280 Malay words of various morphological structures. Of these, data from 1,000 monomorphemic words were analysed in a series of generalised additive mixed models (GAMMs). Root family size and OSC were significant predictors of decision latencies, particularly for lower frequency words. We found a facilitative effect of root family size and OSC. Furthermore, we observed an interaction between root family size and OSC in that an effect of OSC was only apparent in words with larger root families. This interaction has not yet been explored in English but has the potential to be a new benchmark effect to test distributional models of word processing.
RESUMO
LASTU is a tool for searching for Finnish language stimulus words for psycholinguistic studies. The tool allows the user to query a number of properties, including forms, lemmas, frequencies, and morphological features. It also includes two new measures for quantifying lemma and form ambiguity. The tool is written in Python and is available for Windows and macOS platforms. It is available at https://osf.io/j8v6b/ . Included with the tool is a database based on a massive corpus of dependency-parsed Finnish language data crawled from the Internet (over 5 billion tokens). While LASTU has been developed for researchers working on the Finnish language, the openly available implementation can also be applied to other languages.
Assuntos
Idioma , Psicolinguística , Psicolinguística/métodos , Humanos , Finlândia , Software , InternetRESUMO
Sheaves are mathematical objects that describe the globally compatible data associated with open sets of a topological space. Original examples of sheaves were continuous functions; later they also became powerful tools in algebraic geometry, as well as logic and set theory. More recently, sheaves have been applied to the theory of contextuality in quantum mechanics. Whenever the local data are not necessarily compatible, sheaves are replaced by the simpler setting of presheaves. In previous work, we used presheaves to model lexically ambiguous phrases in natural language and identified the order of their disambiguation. In the work presented here, we model syntactic ambiguities and study a phenomenon in human parsing called garden-pathing. It has been shown that the information-theoretic quantity known as 'surprisal' correlates with human reading times in natural language but fails to do so in garden-path sentences. We compute the degree of signalling in our presheaves using probabilities from the large language model BERT and evaluate predictions on two psycholinguistic datasets. Our degree of signalling outperforms surprisal in two ways: (i) it distinguishes between hard and easy garden-path sentences (with a [Formula: see text]-value [Formula: see text]), whereas existing work could not, (ii) its garden-path effect is larger in one of the datasets (32 ms versus 8.75 ms per word), leading to better prediction accuracies. This article is part of the theme issue 'Quantum contextuality, causality and freedom of choice'.
RESUMO
The Large Database of English Pseudo-compounds (LaDEP) contains nearly 7500 English words which mimic, but do not truly possess, a compound morphemic structure. These pseudo-compounds can be parsed into two free morpheme constituents (e.g., car-pet), but neither constituent functions as a morpheme within the overall word structure. The items were manually coded as pseudo-compounds, further coded for features related to their morphological structure (e.g., presence of multiple affixes, as in ruler-ship), and summarized using common psycholinguistic variables (e.g., length, frequency). This paper also presents an example analysis comparing the lexical decision response times between compound words, pseudo-compound words, and monomorphemic words. Pseudo-compounds and monomorphemic words did not differ in response time, and both groups had slower response times than compound words. This analysis replicates the facilitatory effect of compound constituents during lexical processing, and demonstrates the need to emphasize the pseudo-constituent structure of pseudo-compounds to parse their effects. Further applications of LaDEP include both psycholinguistic studies investigating the nature of human word processing or production and educational or clinical settings evaluating the impact of linguistic features on language learning and impairments. Overall, the items within LaDEP provide a varied and representative sample of the population of English pseudo-compounds which may be used to facilitate further research related to morphological decomposition, lexical access, meaning construction, orthographical influences, and much more.
Assuntos
Idioma , Vocabulário , Humanos , Psicolinguística , Linguística , Desenvolvimento da Linguagem , SemânticaRESUMO
Proper names are especially prone to retrieval failures and tip-of-the-tongue states (TOTs)-a phenomenon wherein a person has a strong feeling of knowing a word but cannot retrieve it. Current research provides mixed evidence regarding whether related names facilitate or compete with target-name retrieval. We examined this question in two experiments using a novel paradigm where participants either read a prime name aloud (Experiment 1) or classified a written prime name as famous or non-famous (Experiment 2) prior to naming a celebrity picture. Successful retrievals decreased with increasing trial number (and was dependent on the number of previously presented similar famous people) in both experiments, revealing a form of accumulating interference between multiple famous names. However, trial number had no effect on TOTs, and within each trial famous prime names increased TOTs only in Experiment 2. These results can be explained within a framework that assumes competition for selection at the point of lexical retrieval, such that successful retrievals decrease after successive retrievals of proper names of depicted faces of semantically similar people. By contrast, the effects of written prime words only occur when prime names are sufficiently processed, and do not provide evidence for competition but may reflect improved retrieval relative to a "don't know" response.
Assuntos
Rememoração Mental , Nomes , Humanos , Rememoração Mental/fisiologia , Leitura , LínguaRESUMO
An experiment investigated the "good-enough" processing account regarding how people parse sentences with late-closure ambiguity, such as While Anna dressed the baby who was cute and cuddly spit up on the bed. One possible result of an initial misparse of the sentence (thinking that Anna dressed the baby) is that the correct parse then cannot be created. The alternative is that, although the misparse may linger in the comprehender's mind, the correct parse is eventually established and coexists with the misparse. This study approached this issue through an analogy to quantum physics. When photons are directed toward two small slits, it appears as if each photon passes through both slits ("bothness"). If particles not subject to quantum effects are directed at the slits, they pass through one or the other ("oneness"). Participants read sentences containing the late-closure ambiguity and afterward answered two questions about each sentence. These could query the potential misparse (Did Anna dress the baby?), correct-parse (Did Anna dress herself?), or the main clause (Did the baby spit up on the bed?), and the order of the question types was varied to test for quantum measurement context effects. The results supported "oneness," that is, the correct-parse and misparse of the subordinate clause do not coexist. Participants rarely said "yes" to both the misparse and correct-parse questions, and the "yes" response proportions for these two questions invariably added up to around 1.0. Furthermore, no quantum-like measurement-order effect between misparse and correct-parse questions was found.
Assuntos
Idioma , Leitura , Humanos , Compreensão/fisiologia , PsicolinguísticaRESUMO
Previous literature has indicated conflicting results regarding a response time bias favouring words indicating large real-world objects (RWO) over words indicating small RWO during a lexical decision task. This study aimed to replicate an original experiment and, expanding on it, disentangle possible alternatives for why this effect is sometimes observed and sometimes not. The same methods as the original study were followed, and the results were inconsistent with all previously published findings. Although no significant difference was observed for response time, the findings indicated a significant difference in accuracy and inverse efficiency scores such that "large" words were recognised significantly more accurately than "small" words. After examining several linguistic dimensions that may also contribute to response time, statistical models accounting for these dimensions yielded a significant and increased effect size for the response time size rating of words in our sample from the United States. Our findings indicate that there is a cognitive bias favouring words representing large RWO over small ones but suggest several additional linguistic factors need to be controlled for it to be detected consistently in response time.