Semantic Prediction Research Articles

Verb-based prediction during language processing: the case of Dutch and Turkish

This study investigated whether cross-linguistic differences affect semantic prediction. We assessed this by looking at two languages, Dutch and Turkish, that differ in word order and thus vary in how words come together to create sentence meaning. In an eye-tracking task, Dutch and Turkish four-year-olds (N = 40), five-year-olds (N = 58), and adults (N = 40) were presented with a visual display containing two familiar objects (e.g., a cake and a tree). Participants heard semantically constraining (e.g., "The boy eats the big cake") or neutral sentences (e.g., "The boy sees the big cake") in their native language. The Dutch data revealed a prediction effect for children and adults; however, it was larger for the adults. The Turkish data revealed no prediction effect for the children but only for the adults. These findings reveal that experience with word order structures and/or automatization of language processing routines may lead to timecourse differences in semantic prediction.

Semantic prediction assistant approach applied to energy efficiency in Tertiary buildings

Semantic prediction assistant approach applied to energy efficiency in Tertiary buildings

Semantic weldability prediction with RSW quality dataset and knowledge construction

This paper presents a semantic Resistance Spot Welding (RSW) weldability prediction framework. The framework constructs a shareable weldability knowledge database based on the regression rules from inconsistent RSW quality datasets. This research aims to effectively predict the weldability of RSW process for existing or new weldment design. A real welding test dataset collected from an automotive OEM is used to extract decision rules using a decision tree algorithm, Classification and Regression Trees (CART). The extracted decision rules are converted systematically into SWRL rules for capturing the semantics and to increase the shareability of the constructed knowledge. The experiments show that the RSW ontology, along with SWRL rules that contains weldability rules constructed from the datasets, successfully predicts the weldability (nugget width) values for RSW cases. The predicted nugget width values are found to be in-close proximity of the actual values. This paper shows that semantic prediction framework construes an intelligent way for constructing accurate and transparent predictive models for RSW weldability verification.

Semantic Prediction Using Distributed Rule Mining for Web-Based Big Stock Data

Semantic Prediction Using Distributed Rule Mining for Web-Based Big Stock Data

The development of linguistic prediction: Predictions of sound and meaning in 2- to 5-year-olds

Language processing in adults is facilitated by an expert ability to generate detailed predictions about upcoming words. This may seem like an acquired skill, but some models of language acquisition assume that the ability to predict is a prerequisite for learning. This raises a question: Do children learn to predict, or do they predict to learn? We tested whether children, like adults, can generate expectations about not just the meanings of upcoming words but also their sounds, which would be critical for using prediction to learn about language. In two looking-while-listening experiments, we show that 2-year-olds can generate expectations about meaning based on a determiner (Can you see one…ball/two…ice creams?) but that even children as old as 5 years do not show an adult-like ability to predict the phonology of upcoming words based on a determiner (Can you see a…ball/an…ice cream?). Our results, therefore, suggest that the ability to generate detailed predictions is a late-acquired skill. We argue that prediction might not be the key mechanism driving children’s learning, but that the ability to generate accurate semantic predictions may nevertheless have facilitative effects on language development.

Knowledge acquisition is governed by striatal prediction errors

Discrepancies between expectations and outcomes, or prediction errors, are central to trial-and-error learning based on reward and punishment, and their neurobiological basis is well characterized. It is not known, however, whether the same principles apply to declarative memory systems, such as those supporting semantic learning. Here, we demonstrate with fMRI that the brain parametrically encodes the degree to which new factual information violates expectations based on prior knowledge and beliefs—most prominently in the ventral striatum, and cortical regions supporting declarative memory encoding. These semantic prediction errors determine the extent to which information is incorporated into long-term memory, such that learning is superior when incoming information counters strong incorrect recollections, thereby eliciting large prediction errors. Paradoxically, by the same account, strong accurate recollections are more amenable to being supplanted by misinformation, engendering false memories. These findings highlight a commonality in brain mechanisms and computational rules that govern declarative and nondeclarative learning, traditionally deemed dissociable.

Prediction in a visual language: real-time sentence processing in American Sign Language across development

ABSTRACTPrediction during sign language comprehension may enable signers to integrate linguistic and non-linguistic information within the visual modality. In two eye-tracking experiments, we investigated American Sign language (ASL) semantic prediction in deaf adults and children (aged 4–8 years). Participants viewed ASL sentences in a visual world paradigm in which the sentence-initial verb was either neutral or constrained relative to the sentence-final target noun. Adults and children made anticipatory looks to the target picture before the onset of the target noun in the constrained condition only, showing evidence for semantic prediction. Crucially, signers alternated gaze between the stimulus sign and the target picture only when the sentential object could be predicted from the verb. Signers therefore engage in prediction by optimising visual attention between divided linguistic and referential signals. These patterns suggest that prediction is a modality-independent process, and theoretical implications are discussed.

Scene grammar shapes the way we interact with objects, strengthens memories, and speeds search

Predictions of environmental rules (here referred to as “scene grammar”) can come in different forms: seeing a toilet in a living room would violate semantic predictions, while finding a toilet brush next to the toothpaste would violate syntactic predictions. The existence of such predictions has usually been investigated by showing observers images containing such grammatical violations. Conversely, the generative process of creating an environment according to one’s scene grammar and its effects on behavior and memory has received little attention. In a virtual reality paradigm, we either instructed participants to arrange objects according to their scene grammar or against it. Subsequently, participants’ memory for the arrangements was probed using a surprise recall (Exp1), or repeated search (Exp2) task. As a result, participants’ construction behavior showed strategic use of larger, static objects to anchor the location of smaller objects which are generally the goals of everyday actions. Further analysis of this scene construction data revealed possible commonalities between the rules governing word usage in language and object usage in naturalistic environments. Taken together, we revealed some of the building blocks of scene grammar necessary for efficient behavior, which differentially influence how we interact with objects and what we remember about scenes.

Adjectival inflection as diagnostic for structural position: inside and outside the Icelandic definiteness domain

This article attempts to account for the distribution of Icelandic adjectival inflection in a manner that also captures a problematic case that has not been satisfactorily analyzed in the literature. It is argued that weak inflection is triggered if the adjective is c-commanded by a feature [definite]. Strong inflection occurs precisely if weak inflection is not triggered. This implementation accounts for the occurrence of strong inflection on predicative adjectives and adnominal adjectives in indefinite noun phrases, but moreover allows us to account for an unexpected pattern in Icelandic where a strongly inflected adjective occurs in a definite noun phrase. It is argued that, in this case, the adjective is not c-commanded by the definite article, but, in fact, merged outside DP. This rather unorthodox assumption motivated on morpho-syntactic grounds makes a number of syntactic and semantic predictions. In particular, adjectives that are merged outside DP are expected to modify not simply an NP (of type $${ }$$ ), but actually a referential expression/an individual of type e. The discussion of four instances of that strongly inflected pattern—appositives, expressives, positional predicates and little partitives—provides various kinds of evidence and shows that these expectations are indeed borne out.

Semantic knowledge and action-based visual prediction

Semantic knowledge and action-based visual prediction

Facilitating morphosyntactic and semantic prediction among second language speakers of German

ABSTRACTUsing two visual priming experiments, the present study investigates whether presenting facilitative semantic (i.e. colour) and morphosyntactic (i.e. grammatical gender) information in a prime image prior to reading a target sentence facilitates naming of a sentence-final target image among L1 German and L1 English-L2 German speakers. In Experiment 1, L1 and L2 German speakers used both semantic and gender cues to predict the sentence-final target image. In Experiment 2, a new group of L2 German speakers used gender cues to predict, but this effect was stronger when gender information was provided via a gender-marked indefinite article and adjective in the prime, than when the prime contained only the gender-marked article. These results suggest that if L2 speakers are able to overcome unstable and often inaccurate L2 gender assignment, they can use gender cues in a native-like manner for prediction, but that multiple gender-marked cues may be necessary for such prediction to occur.

Why should we care about sound symbolism in EFL learning?: Two pilot studies

Given the importance of the phonological and lexical components of the language in L2 learning, this article discusses an innovative, holistic approach to learning these two components of the language based on the existence of “sound symbolism”—the interrelation between sound and meaning—in English. In particular, it describes how and why the study of sound symbolism can be advantageous to EFL learners. This claim is grounded in empirical data gathered from two pilot studies carried out in two educational settings (a secondary school and the University of Alicante). The results suggest that knowledge of sound symbolic principles underlying the English language can enhance lexical storage and semantic prediction

Predicting semantic features in Chinese: Evidence from ERPs

This article reports two ERP studies that exploited the classifier system of Mandarin Chinese to investigate semantic prediction. In Mandarin, in certain contexts, a noun has to be preceded by a classifier, which has to match the noun in semantically-defined features. In both experiments, an N400 effect was elicited in response to a classifier that mismatched an up-coming predictable noun, relative to a matching classifier. Among the mismatching classifiers, the N400 effect was graded, being smaller for classifiers that were semantically related to the predicted word, relative to classifiers that were semantically unrelated to the predicted word. Given that the classifier occurred before the predicted word, this result shows that fine-grained semantic features of nouns can be pre-activated in advance of bottom-up input. The studies thus extend previous findings based on a more restricted range of highly grammaticalized features such as gender or animacy in Indo-European languages (Szewczyk & Schriefers, 2013; Van Berkum, Brown, Zwitserlood, Kooijman, & Hagoort, 2005; Wicha, Bates, Moreno, & Kutas, 2003).

Pace Yourself: Intraindividual Variability in Context Use Revealed by Self-paced Event-related Brain Potentials.

Event-related brain potentials (ERPs) have revealed multiple mechanisms by which contextual constraints impact language processing. At the same time, little work has examined the trial-to-trial dynamics of context use in the brain. In the current study, we probed intraindividual variability in behavioral and neural indices of context processing during reading. In a concurrent self-paced reading and ERP paradigm, participants read sentences that were either strongly or weakly constraining completed with an expected or unexpected target word. Our findings revealed substantial within-subject variability in behavioral and neural responses to contextual constraints. First, context-based amplitude reductions of the N400, a component linked to semantic memory access, were largest among trials eliciting the slowest RTs. Second, the RT distribution of unexpected words in strongly constraining contexts was positively skewed, reflecting an increased proportion of very slow RTs to trials that violated semantic predictions. Among those prediction-violating trials eliciting faster RTs, a late sustained anterior positivity was observed. However, among trials producing the differentially slowed RTs to prediction violations, we observed a markedly earlier effect of constraint in the form of an anterior N2, a component linked to conflict resolution and the cognitive control of behavior. The current study provides the first neurophysiological evidence for the direct role of cognitive control functions in the volitional control of reading. Collectively, our findings suggest that context use varies substantially within individual participants and that coregistering behavioral and neural indices of online sentence processing offers a window into these single-item dynamics.

Neural Correlates of Semantic Prediction and Resolution in Sentence Processing.

Most brain-imaging studies of language comprehension focus on activity following meaningful stimuli. Testing adult human participants with high-density EEG, we show that, already before the presentation of a critical word, context-induced semantic predictions are reflected by a neurophysiological index, which we therefore call the semantic readiness potential (SRP). The SRP precedes critical words if a previous sentence context constrains the upcoming semantic content (high-constraint contexts), but not in unpredictable (low-constraint) contexts. Specific semantic predictions were indexed by SRP sources within the motor system—in dorsolateral hand motor areas for expected hand-related words (e.g., “write”), but in ventral motor cortex for face-related words (“talk”). Compared with affirmative sentences, negated ones led to medial prefrontal and more widespread motor source activation, the latter being consistent with predictive semantic computation of alternatives to the negated expected concept. Predictive processing of semantic alternatives in negated sentences is further supported by a negative-going event-related potential at ∼400 ms (N400), which showed the typical enhancement to semantically incongruent sentence endings only in high-constraint affirmative contexts, but not to high-constraint negated ones. These brain dynamics reveal the interplay between semantic prediction and resolution (match vs error) processing in sentence understanding.SIGNIFICANCE STATEMENT Most neuroscientists agree on the eminent importance of predictive mechanisms for understanding basic as well as higher brain functions. This contrasts with a sparseness of brain measures that directly reflects specific aspects of prediction, as they are relevant in the processing of language and thought. Here we show that when critical words are strongly expected in their sentence context, a predictive brain response reflects meaning features of these anticipated symbols already before they appear. The granularity of the semantic predictions was so fine grained that the cortical sources in sensorimotor and medial prefrontal cortex even distinguished between predicted face- or hand-related action words (e.g., the words “lick” or “pick”) and between affirmative and negated sentence meanings.

Cerebellar tDCS Modulates Neural Circuits during Semantic Prediction: A Combined tDCS-fMRI Study.

It has been proposed that the cerebellum acquires internal models of mental processes that enable prediction, allowing for the optimization of behavior. In language, semantic prediction speeds speech production and comprehension. Right cerebellar lobules VI and VII (including Crus I/II) are engaged during a variety of language processes and are functionally connected with cerebral cortical language networks. Further, right posterolateral cerebellar neuromodulation modifies behavior during predictive language processing. These data are consistent with a role for the cerebellum in semantic processing and semantic prediction. We combined transcranial direct current stimulation (tDCS) and fMRI to assess the behavioral and neural consequences of cerebellar tDCS during a sentence completion task. Task-based and resting-state fMRI data were acquired in healthy human adults (n = 32; μ = 23.1 years) both before and after 20 min of 1.5 mA anodal (n = 18) or sham (n = 14) tDCS applied to the right posterolateral cerebellum. In the sentence completion task, the first four words of the sentence modulated the predictability of the final target word. In some sentences, the preceding context strongly predicted the target word, whereas other sentences were nonpredictive. Completion of predictive sentences increased activation in right Crus I/II of the cerebellum. Relative to sham tDCS, anodal tDCS increased activation in right Crus I/II during semantic prediction and enhanced resting-state functional connectivity between hubs of the reading/language networks. These results are consistent with a role for the right posterolateral cerebellum beyond motor aspects of language, and suggest that cerebellar internal models of linguistic stimuli support semantic prediction.SIGNIFICANCE STATEMENT Cerebellar involvement in language tasks and language networks is now well established, yet the specific cerebellar contribution to language processing remains unclear. It is thought that the cerebellum acquires internal models of mental processes that enable prediction, allowing for the optimization of behavior. Here we combined neuroimaging and neuromodulation to provide evidence that the cerebellum is specifically involved in semantic prediction during sentence processing. We found that activation within right Crus I/II was enhanced when semantic predictions were made, and we show that modulation of this region with transcranial direct current stimulation alters both activation patterns and functional connectivity within whole-brain language networks. For the first time, these data show that cerebellar neuromodulation impacts activation patterns specifically during predictive language processing.

Modeling the N400 ERP component as transient semantic over-activation within a neural network model of word comprehension

The study of the N400 event-related brain potential has provided fundamental insights into the nature of real-time comprehension processes, and its amplitude is modulated by a wide variety of stimulus and context factors. It is generally thought to reflect the difficulty of semantic access, but formulating a precise characterization of this process has proved difficult. Laszlo and colleagues (Laszlo & Plaut, 2012; Laszlo & Armstrong, 2014) used physiologically constrained neural networks to model the N400 as transient over-activation within semantic representations, arising as a consequence of the distribution of excitation and inhibition within and between cortical areas. The current work extends this approach to successfully model effects on both N400 amplitudes and behavior of word frequency, semantic richness, repetition, semantic and associative priming, and orthographic neighborhood size. The account is argued to be preferable to one based on “implicit semantic prediction error” (Rabovsky & McRae, 2014) for a number of reasons, the most fundamental of which is that the current model actually produces N400-like waveforms in its real-time activation dynamics.

Prediction Signatures in the Brain: Semantic Pre-Activation during Language Comprehension

There is broad agreement that context-based predictions facilitate lexical-semantic processing. A robust index of semantic prediction during language comprehension is an evoked response, known as the N400, whose amplitude is modulated as a function of semantic context. However, the underlying neural mechanisms that utilize relations of the prior context and the embedded word within it are largely unknown. We measured magnetoencephalography (MEG) data while participants were listening to simple German sentences in which the verbs were either highly predictive for the occurrence of a particular noun (i.e., provided context) or not. The identical set of nouns was presented in both conditions. Hence, differences for the evoked responses of the nouns can only be due to differences in the earlier context. We observed a reduction of the N400 response for highly predicted nouns. Interestingly, the opposite pattern was observed for the preceding verbs: highly predictive (that is more informative) verbs yielded stronger neural magnitude compared to less predictive verbs. A negative correlation between the N400 effect of the verb and that of the noun was found in a distributed brain network, indicating an integral relation between the predictive power of the verb and the processing of the subsequent noun. This network consisted of left hemispheric superior and middle temporal areas and a subcortical area; the parahippocampus. Enhanced activity for highly predictive relative to less predictive verbs, likely reflects establishing semantic features associated with the expected nouns, that is a pre-activation of the expected nouns.

Predictive Brain Mechanisms in Sound-to-Meaning Mapping during Speech Processing.

In daily communication via spoken language, one of the core processes is understanding the words being used. Effortless and efficient information exchange via speech relies not only on the identification of individual spoken words, but also on the contextual information giving rise to expected meanings. Despite the accumulating evidence for the bottom-up perception of auditory input, it is still not fully understood how the top-down modulation is achieved in the extensive frontotemporal cortical network. Here, we provide a comprehensive description of the neural substrates underlying sound-to-meaning mapping and demonstrate how the dual-stream model functions in the modulation of expectations, allowing for a better understanding of how the human brain accomplishes sound-to-meaning mapping in a predictive manner.

Semantic prediction in language comprehension: evidence from brain potentials

ABSTRACTDo people predict specific word-forms during language comprehension? In an Event-Related Potential (ERP) study participants read German sentences with predictable (The goalkeeper claims that the slick ball was easy to CATCH.) and unpredictable (The kids boasted that the young horse was easy to SADDLE.) verbs. Verbs were either consistent with the expected word-form (catch/saddle) or inconsistent and therefore led to ungrammaticality (*catches/*saddles). ERPs within the N400 time-window were modulated by predictability but not by the surface-form of the verbs, suggesting that no exact word-forms were predicted. Based on our results we will argue that predictions included semantic rather than form-information. Furthermore, ungrammatical verbs led to a strong P600, probably due to task-saliency whereas correct unpredictable verbs elicited an anterior post-N400 positivity. Because the contexts were moderately constraining, this might reflect discourse revision processes rather than inhibition of a predicted word.