Periodic Reporting for period 1 - BraveNewWord (The acquisition of new meanings through novel word learning)
Période du rapport: 2023-06-01 au 2025-11-30
Across the whole lifespan, lexical acquisition is a process always at place. Indeed, we learn a new word every other day, even if the word referent is not present in the environment and no explicit instruction concerning the word meaning or usage is provided. Often we do not even realize that the word we are experiencing is, in fact, unfamiliar. The reason novel word acquisition is so seamlessly achieved is that, with novel words, we also typically acquire novel meanings: such parallel semantic enrichment makes the newly encountered word meaningful, and can be considered the very purpose of lexical acquisition. Despite its pervasiveness, such a remarkable phenomenon is rarely studied in adults and, as a result, poorly understood. BraveNewWord aims at filling this gap by bringing together approaches from computational linguistics to quantitatively formalize the semantic side of novel word acquisition, and methods from experimental psychology and cognitive neuroscience to validate the predictions of the resulting models in a cognitive perspective.
BraveNewWord posits three main cognitive mechanisms of semantic enrichment via novel words. First, the novel meaning can be induced via the sentence context in which the new word appears. Indeed, most novel words are not found in isolation, and the surrounding linguistic environment is informative about their meaning: when we hear “The small, hairy wug was sleeping below a tree” we have an intuition that the unfamiliar “wug” might denote some kind of animal. Second, substantial information can be obtained on the basis of the word morphological structure; in fact, most of the novel words are composed by familiar sublexical units, i.e. morphemes: the meaning of “quickify” is immediately evident since we are familiar with both “quick” and “ify”. Third, natural languages present a certain degree of systematicity, i.e. nuanced association between form and meaning; such reliable statistical patterns can be exploited to have an intuition about a novel word, even if it does not include any particularly salient or familiar sublexical element (e.g. futmaw).
In BraveNewWord, the three described mechanisms are integrated in a unique computational framework to provide a new understanding of the relation between language and meaning, and the cognitive underpinnings on which such relation is built.
BraveNewWord posits three main cognitive mechanisms of semantic enrichment via novel words. First, the novel meaning can be induced via the sentence context in which the new word appears. Indeed, most novel words are not found in isolation, and the surrounding linguistic environment is informative about their meaning: when we hear “The small, hairy wug was sleeping below a tree” we have an intuition that the unfamiliar “wug” might denote some kind of animal. Second, substantial information can be obtained on the basis of the word morphological structure; in fact, most of the novel words are composed by familiar sublexical units, i.e. morphemes: the meaning of “quickify” is immediately evident since we are familiar with both “quick” and “ify”. Third, natural languages present a certain degree of systematicity, i.e. nuanced association between form and meaning; such reliable statistical patterns can be exploited to have an intuition about a novel word, even if it does not include any particularly salient or familiar sublexical element (e.g. futmaw).
In BraveNewWord, the three described mechanisms are integrated in a unique computational framework to provide a new understanding of the relation between language and meaning, and the cognitive underpinnings on which such relation is built.
Initially, BraveNewWord focused on defining its computational framework, implementing computational characterizations of the three mechanisms of novel-word-driven meaning acquisition in the domain of vector space modelling. Such a view characterizes meaning as numerical vectors, which can in turn be seen as points in a multidimensional space. When such a semantic space is aligned with a linguistic space, populated by words and sublexical elements, one can predict which meaning will be evoked by novel, unfamiliar elements. The computational mechanisms allowing for such predictions are grounded in the three cognitive mechanisms described above: meaning induction from minimal linguistic context, semantic combination of morphological units, form-to-meaning mapping. To achieve its computational objectives, BraveNewWord applies both existing architectures and newly developed approaches.
The models developed by BraveNewWord naturally produce quantitative, empirically testable predictions about behaviour and neural activity. The project is testing such predictions with methodologies ranging from response times to neuroimaging to electrophysiology.
Concerning the impact of minimal linguistic context, we have observed a modulation of the N400 for novel words in context, after as few as two occurrences. The N400 is an electrophysiological response, measured via EEG, that indexes how surprising a word is within a given sentence or, from a different perspective, how difficult it is to integrate the encountered element in the previous context. This evidence indicates that novel words are rapidly assigned a meaning, which is routinely integrated with the previously presented familiar information. Crucially, the BraveNewWord computational approach can predict the N400 magnitude, and hence how well the novel word integrates with the preceding context.
Concerning morphology-induced meanings, we relied on and extended a model, CAOSS, previously proposed for novel compound words (e.g. rivercat). We adapted this architecture to other types of morphologically complex elements, such as prefixed (e.g. respeak) and suffixed words (e.g. quickify), and showed that the model predictions align with human responses in behavioral tasks across different languages. Furthermore, in a neuroimaging study, we observed distinct neural signatures for novel meanings induced by morphologically complex words.
Concerning form-meaning mapping, across a number of studies we investigated human intuitions about the possible meaning of completely unfamiliar linguistic strings (e.g. futmaw). Participants were shown to be able to produce consistent responses across a range of different semantic dimensions, up to actual definitions. Such responses were significantly predicted by the BraveNewWord computational approach. Moreover, uniquely within the BraveNewWord endeavour that typically moves from language to semantics, we designed a system that moves in the opposite direction, producing a new word on the basis of a desired meaning.
The models developed by BraveNewWord naturally produce quantitative, empirically testable predictions about behaviour and neural activity. The project is testing such predictions with methodologies ranging from response times to neuroimaging to electrophysiology.
Concerning the impact of minimal linguistic context, we have observed a modulation of the N400 for novel words in context, after as few as two occurrences. The N400 is an electrophysiological response, measured via EEG, that indexes how surprising a word is within a given sentence or, from a different perspective, how difficult it is to integrate the encountered element in the previous context. This evidence indicates that novel words are rapidly assigned a meaning, which is routinely integrated with the previously presented familiar information. Crucially, the BraveNewWord computational approach can predict the N400 magnitude, and hence how well the novel word integrates with the preceding context.
Concerning morphology-induced meanings, we relied on and extended a model, CAOSS, previously proposed for novel compound words (e.g. rivercat). We adapted this architecture to other types of morphologically complex elements, such as prefixed (e.g. respeak) and suffixed words (e.g. quickify), and showed that the model predictions align with human responses in behavioral tasks across different languages. Furthermore, in a neuroimaging study, we observed distinct neural signatures for novel meanings induced by morphologically complex words.
Concerning form-meaning mapping, across a number of studies we investigated human intuitions about the possible meaning of completely unfamiliar linguistic strings (e.g. futmaw). Participants were shown to be able to produce consistent responses across a range of different semantic dimensions, up to actual definitions. Such responses were significantly predicted by the BraveNewWord computational approach. Moreover, uniquely within the BraveNewWord endeavour that typically moves from language to semantics, we designed a system that moves in the opposite direction, producing a new word on the basis of a desired meaning.
BraveNewWord is proposing a comprehensive framework to explain the semantic counterpart of novel word acquisition. The project is showing how meaningful the processing of unfamiliar linguistic elements can be, indicating that semantic memory is substantially active even when there is no established link between a symbol and the meaning it connotes. This uncovers new dynamics in the human mind, blurring the boundaries between language and semantic memory and potentially challenging the very notion of “word” as a psychologically-defined category.
Given its computationally-driven approach, BraveNewWord will further pave the ground for new applications and research venues. It will allow us to understand which novel elements are easier to integrate in memory, helping with data-driven definitions of education programs in second language acquisition and interventions in the rehabilitation field. Moreover, BraveNewWord can tell us which words can be more impactful in shaping ideas, and hence nudge behaviors. This has applications in marketing (e.g. brand generation), but also in promoting desirable societal change, such as in the domain of inclusive language
Given its computationally-driven approach, BraveNewWord will further pave the ground for new applications and research venues. It will allow us to understand which novel elements are easier to integrate in memory, helping with data-driven definitions of education programs in second language acquisition and interventions in the rehabilitation field. Moreover, BraveNewWord can tell us which words can be more impactful in shaping ideas, and hence nudge behaviors. This has applications in marketing (e.g. brand generation), but also in promoting desirable societal change, such as in the domain of inclusive language