Spreading Activation in an Attractor Network With Latching Dynamics: Automatic Semantic Priming Revisited

Localist models of spreading activation (SA) and models assuming distributed representations offer very different takes on semantic priming, a widely investigated paradigm in word recognition and semantic memory research. In this study, we implemented SA in an attractor neural network model with distributed representations and created a unified framework for the two approaches. Our models assume a synaptic depression mechanism leading to autonomous transitions between encoded memory patterns (latching dynamics), which account for the major characteristics of automatic semantic priming in humans. Using computer simulations, we demonstrated how findings that challenged attractor‐based networks in the past, such as mediated and asymmetric priming, are a natural consequence of our present model’s dynamics. Puzzling results regarding backward priming were also given a straightforward explanation. In addition, the current model addresses some of the differences between semantic and associative relatedness and explains how these differences interact with stimulus onset asynchrony in priming experiments.     

The Social Representations of the Bali Climate Conference in the French and German Media

Climate change is a major current affair for which recent United Nations climate conferences aim to build consensus and develop international solutions. The objective of this article is to compare, through the theoretical lens of social representations, the way in which French and German media, specifically newspapers, represent the Bali climate conference. We use the triangulation of data analysis to take both the pragmatic and the semantic aspects of media discourse into account. Results show that German media adopt both a local and a global vision of climate change and of the conference. Religious metaphors highlight a moral dimension of the conference, suggesting anchoring in human and political categories. In contrast, in French media, we identify that conflicts between countries render the stakes of climate change concrete by war metaphors. The French discourses examined are shown to be organised through the anchoring of political and financial categories. Results are discussed in relation to the history of green movements in the two countries and in relation to practical implications. Copyright © 2011 John Wiley & Sons, Ltd.     

Basic Human Values and Their Contexts: A Multilevel Analysis of Support for the Use of Armed Drones in the United States,United Kingdom,and Turkey

Armed drones are now a key component of military strategy; however, little empirical research has explored the phenomenon in terms of psychological processes. Previous research has emphasized the importance of basic human values for structuring understandings of and opinions towards foreign policy issues. Using a social representations approach, we used a multilevel modeling approach to explore the link between values and support for the use of armed drones in the United Kingdom, the United States, and Turkey. In line with our predictions, high priority of self-transcendence values negatively predicted, and high priority of conservation values positively predicted, support for armed drones. Furthermore, given our theoretical framework, we specified that values should be conceptualized as prioritized or devalued within a particular context, and when values were specified as country-level, as well as individual predictors, this led to an increase in model fit. These findings are discussed in light of a developing line of research on meta-representations and their consequences for political opinion, and directions for future research are advanced.     

Extracting Low-Dimensional Psychological Representations from Convolutional Neural Networks

Convolutional neural networks (CNNs) are increasingly widely used in psychology and neuroscience to predict how human minds and brains respond to visual images. Typically, CNNs represent these images using thousands of features that are learned through extensive training on image datasets. This raises a question: How many of these features are really needed to model human behavior? Here, we attempt to estimate the number of dimensions in CNN representations that are required to capture human psychological representations in two ways: (1) directly, using human similarity judgments and (2) indirectly, in the context of categorization. In both cases, we find that low-dimensional projections of CNN representations are sufficient to predict human behavior. We show that these low-dimensional representations can be easily interpreted, providing further insight into how people represent visual information. A series of control studies indicate that these findings are not due to the size of the dataset we used and may be due to a high level of redundancy in the features appearing in CNN representations.     

Reuse and body-formatted representations in simulation theory

I review the shifting definitions of simulation as found in recent versions of the simulation theory (ST) of social cognition. I focus on two concepts that have become central to recent ST in the work of a number of simulation theorists: the notion of reuse and the notion of B-formatted representations. I point out specific limitations or problems involved in these concepts. Although the reuse hypothesis provides an interesting evolutionary account of how neural mechanisms may adapt to new tasks, it doesn’t offer an explanation of how these mechanisms work. In contrast to the genuinely embodied account that simulation theorists seek, an explanation of social cognition in terms of B-formatted representations not only remains disembodied, but also ignores social interaction and remains solipsistic. I conclude by briefly outlining a non-simulationist enactivist account that can incorporate the reuse hypothesis.     

Modelling the effects of semantic ambiguity in word recognition

Most words in English are ambiguous between different interpretations; words can mean different things in different contexts. We investigate the implications of different types of semantic ambiguity for connectionist models of word recognition. We present a model in which there is competition to activate distributed semantic representations. The model performs well on the task of retrieving the different meanings of ambiguous words, and is able to simulate data reported by Rodd, Gaskell, and Marslen-Wilson [J. Mem. Lang. 46 (2002) 245] on how semantic ambiguity affects lexical decision performance. In particular, the network shows a disadvantage for words with multiple unrelated meanings (e.g., bark) that coexists with a benefit for words with multiple related word senses (e.g., twist). The ambiguity disadvantage arises because of interference between the different meanings, while the sense benefit arises because of differences in the structure of the attractor basins formed during learning. Words with few senses develop deep, narrow attractor basins, while words with many senses develop shallow, broad basins. We conclude that the mental representations of word meanings can be modelled as stable states within a high-dimensional semantic space, and that variations in the meanings of words shape the landscape of this space.