A growing body of evidence suggests that cognition is embodied and grounded. Abstract concepts, though, remain a significant theoretical chal- lenge. A number of researchers have proposed that language makes an important contribution to our capacity to acquire and employ concepts, particularly abstract ones. In this essay, I critically examine this suggestion and ultimately defend a version of it. I argue that a successful account of how language augments cognition should emphasize its symbolic properties and incorporate a view of embodiment (...) that recognizes the flexible, multi- modal and task-related nature of action, emotion and perception systems. On this view, language is an ontogenetically disruptive cognitive technology that expands our conceptual reach. (shrink)

As embodied theories of cognition are increasingly formalized and tested, care must be taken to make informed assumptions regarding the nature of concepts and representations. In this study, we outline three reasons why one cannot, in effect, represent the same concept twice. First, online perception affects offline representation: Current representational content depends on how ongoing demands direct attention to modality-specific systems. Second, language is a fundamental facilitator of offline representation: Bootstrapping and shortcuts within the computationally cheaper linguistic system continuously modify (...) representational content. Third, time itself is a source of representational change: As the content of underlying concepts shifts with the accumulation of direct and vicarious experience, so too does the content of representations that draw upon these concepts. We discuss the ramifications of these principles for research into both human and synthetic cognitive systems. (shrink)

One of the main limitations of natural language‐based approaches to meaning is that they do not incorporate multimodal representations the way humans do. In this study, we evaluate how well different kinds of models account for people's representations of both concrete and abstract concepts. The models we compare include unimodal distributional linguistic models as well as multimodal models which combine linguistic with perceptual or affective information. There are two types of linguistic models: those based on text corpora and those derived (...) from word association data. We present two new studies and a reanalysis of a series of previous studies. The studies demonstrate that both visual and affective multimodal models better capture behavior that reflects human representations than unimodal linguistic models. The size of the multimodal advantage depends on the nature of semantic representations involved, and it is especially pronounced for basic‐level concepts that belong to the same superordinate category. Additional visual and affective features improve the accuracy of linguistic models based on text corpora more than those based on word associations; this suggests systematic qualitative differences between what information is encoded in natural language versus what information is reflected in word associations. Altogether, our work presents new evidence that multimodal information is important for capturing both abstract and concrete words and that fully representing word meaning requires more than purely linguistic information. Implications for both embodied and distributional views of semantic representation are discussed. (shrink)

Over the past 15 years, there have been two increasingly popular approaches to the study of meaning in cognitive science. One, based on theories of embodied cognition, treats meaning as a simulation of perceptual and motor states. An alternative approach treats meaning as a consequence of the statistical distribution of words across spoken and written language. On the surface, these appear to be opposing scientific paradigms. In this review, we aim to show how recent cross-disciplinary developments have done much to (...) reconcile these two approaches. The foundation to these developments has been the recognition that intralinguistic distributional and sensory-motor data are interdependent. We describe recent work in philosophy, psychology, cognitive neuroscience, and computational modeling that are all based on or consistent with this conclusion. We conclude by considering some possible directions for future research that arise as a consequence of these developments. (shrink)

The human conceptual system comprises simulated information of sensorimotor experience and linguistic distributional information of how words are used in language. Moreover, the linguistic shortcut hypothesis predicts that people will use computationally cheaper linguistic distributional information where it is sufficient to inform a task response. In a pre‐registered category production study, we asked participants to verbally name members of concrete and abstract categories and tested whether performance could be predicted by a novel measure of sensorimotor similarity (based on an 11‐dimensional (...) representation of sensorimotor strength) and linguistic proximity (based on word co‐occurrence derived from a large corpus). As predicted, both measures predicted the order and frequency of category production but, critically, linguistic proximity had an effect above and beyond sensorimotor similarity. A follow‐up study using typicality ratings as an additional predictor found that typicality was often the strongest predictor of category production variables, but it did not subsume sensorimotor and linguistic effects. Finally, we created a novel, fully grounded computational model of conceptual activation during category production, which best approximated typical human performance when conceptual activation was allowed to spread indirectly between concepts, and when candidate category members came from both sensorimotor and linguistic distributional representations. Critically, model performance was indistinguishable from typical human performance. Results support the linguistic shortcut hypothesis in semantic processing and provide strong evidence that both linguistic and grounded representations are inherent to the functioning of the conceptual system. All materials, data, and code are available at https://osf.io/vaq56/. (shrink)

Affective information can be retrieved simply by measuring words co‐occurrences in linguistic contexts. Lenci and colleagues demonstrate that the affective measures retrieved from linguistic occurrences predict words’ concreteness: abstract words are more heavily loaded with affective information than concrete ones. These results challenge the Affective grounding hypothesis, suggesting that abstract concepts may be ungrounded and coded only linguistically, and that their affective load may be a linguistic factor.

Theories of embodied cognition assume that concepts are grounded in non-linguistic, sensorimotor experience. In support of this assumption, previous studies have shown that upwards response movements are faster than downwards movements after participants have been presented with words whose referents are typically located in the upper vertical space. This is taken as evidence that processing these words reactivates sensorimotor experiential traces. This congruency effect was also found for novel words, after participants learned these words as labels for novel objects that (...) they encountered either in their upper or lower visual field. While this indicates that direct experience with a word's referent is sufficient to evoke said congruency effects, the present study investigates whether this direct experience is also a necessary condition. To this end, we conducted five experiments in which participants learned novel words from purely linguistic input: Novel words were presented in pairs with real up- or down-words ; they were presented in natural sentences where they replaced these real words ; they were presented as new labels for these real words ; and they were presented as labels for novel combined concepts based on these real words. In all five experiments, we did not find any congruency effects elicited by the novel words; however, participants were always able to make correct explicit judgements about the vertical dimension associated to the novel words. These results suggest that direct experience is necessary for reactivating experiential traces, but this reactivation is not a necessary condition for understanding the corresponding aspects of word meaning. (shrink)

Computational models have shown that purely statistical knowledge about words’ linguistic contexts is sufficient to learn many properties of words, including syntactic and semantic category. For example, models can infer that “postman” and “mailman” are semantically similar because they have quantitatively similar patterns of association with other words. In contrast to these computational results, artificial language learning experiments suggest that distributional statistics alone do not facilitate learning of linguistic categories. However, experiments in this paradigm expose participants to entirely novel words, (...) whereas real language learners encounter input that contains some known words that are semantically organized. In three experiments, we show that the presence of familiar semantic reference points facilitates distributional learning and this effect crucially depends both on the presence of known words and the adherence of these known words to some semantic organization. (shrink)

We compared the ability of three different contextual models of lexical semantic memory and of a simple associative model to predict the properties of semantic networks derived from word association norms. None of the semantic models were able to accurately predict all of the network properties. All three contextual models over-predicted clustering in the norms, whereas the associative model under-predicted clustering. Only a hybrid model that assumed that some of the responses were based on a contextual model and others on (...) an associative network successfully predicted all of the network properties and predicted a word's top five associates as well as or better than the better of the two constituent models. The results suggest that participants switch between a contextual representation and an associative network when generating free associations. We discuss the role that each of these representations may play in lexical semantic memory. Concordant with recent multicomponent theories of semantic memory, the associative network may encode coordinate relations between concepts, and contextual representations may be used to process information about more abstract concepts. (shrink)

This article correlates fine-grained semantic variability and change with measures of occurrence frequency to investigate whether a word’s degree of semantic change is sensitive to how often it is used. We show that this sensitivity can be detected within a short time span (i.e., 20 years), basing our analysis on a large corpus of German allowing for a high temporal resolution (i.e., per month). We measure semantic variability and change with the help of local semantic networks, combining elements of deep (...) learning methodology and graph theory. Our micro-scale analysis complements previous macro-scale studies from the field of natural language processing, corroborating the finding that high token frequency has a negative effect on the degree of semantic change in a lexical item. We relate this relationship to the role of exemplars for establishing form–function pairings between words and their habitual usage contexts. (shrink)

Distributional models of lexical semantics are capable of acquiring sophisticated representations of word meanings. The main theoretical insight provided by these models is that they demonstrate the systematic connection between the knowledge that people acquire and the experience that they have with the natural language environment. However, linguistic experience is inherently variable and differs radically across people due to demographic and cultural variables. Recently, distributional models have been used to examine how word meanings vary across languages and it was found (...) that there is considerable variability in the meanings of words across languages for most semantic categories. The goal of this article is to examine how variable word meanings are across individual language users within a single language. This was accomplished by assembling 500 individual user corpora attained from the online forum Reddit. Each user corpus ranged between 3.8 and 32.3 million words each, and a count‐based distributional framework was used to extract word meanings for each user. These representations were then used to estimate the semantic alignment of word meanings across individual language users. It was found that there are significant levels of relativity in word meanings across individuals, and these differences are partially explained by other psycholinguistic factors, such as concreteness, semantic diversity, and social aspects of language usage. These results point to word meanings being fundamentally relative and contextually fluid, with this relativeness being related to the individualized nature of linguistic experience. (shrink)

Traditional methods for deriving property-based representations of concepts from text have focused on either extracting only a subset of possible relation types, such as hyponymy/hypernymy (e.g., car is-a vehicle) or meronymy/metonymy (e.g., car has wheels), or unspecified relations (e.g., car—petrol). We propose a system for the challenging task of automatic, large-scale acquisition of unconstrained, human-like property norms from large text corpora, and discuss the theoretical implications of such a system. We employ syntactic, semantic, and encyclopedic information to guide our extraction, (...) yielding concept-relation-feature triples (e.g., car be fast, car require petrol, car cause pollution), which approximate property-based conceptual representations. Our novel method extracts candidate triples from parsed corpora (Wikipedia and the British National Corpus) using syntactically and grammatically motivated rules, then reweights triples with a linear combination of their frequency and four statistical metrics. We assess our system output in three ways: lexical comparison with norms derived from human-generated property norm data, direct evaluation by four human judges, and a semantic distance comparison with both WordNet similarity data and human-judged concept similarity ratings. Our system offers a viable and performant method of plausible triple extraction: Our lexical comparison shows comparable performance to the current state-of-the-art, while subsequent evaluations exhibit the human-like character of our generated properties. (shrink)