The use theory of meaning and semantic stipulation

According to Horwich’s use theory of meaning, the meaning of a word W is engendered by the underived acceptance of certain sentences containing W. Horwich applies this theory to provide an account of semantic stipulation: Semantic stipulation proceeds by deciding to accept sentences containing an as yet meaningless word W. Thereby one brings it about that W gets an underived acceptance property. Since a word’s meaning is constituted by its (basic) underived acceptance property, this decision endows the word with a meaning. The use-theoretic account of semantic stipulation contrasts with the standard view that semantic stipulation proceeds by assigning the meaning (reference) to W that makes a certain set of sentences express true propositions. In this paper I will argue that the use-theoretic account does not work. I take Frege to have already made the crucial point: "a definition does not assert anything but lays down something ["etwas festsetzt"]” (Frege 1899, 36). A semantic stipulation for W cannot be the decision to accept a sentence containing W or be explained in terms of such an acceptance. Semantic stipulation constitutes a problem for Horwich's use theory of meaning, especially his basic notion of acceptance.

Context,content, and relativism

This paper argues against relativism, focusing on relativism based on the semantics of predicates of personal taste. It presents and defends a contextualist semantics for these predicates, derived from current work on gradable adjectives. It then considers metasemantic questions about the kinds of contextual parameters this semantics requires. It argues they are not metasemantically different from those in other gradable adjectives, and that contextual parameters of this sort are widespread in natural language. Furthermore, this paper shows that if such parameters are rejected, it leads to an unacceptably rampant form of relativism, that relativizes truth to an open-ended list of parameters.

Conceptual Hierarchies in a Flat Attractor Network: Dynamics of Learning and Computations

The structure of people's conceptual knowledge of concrete nouns has traditionally been viewed as hierarchical ( Collins & Quillian, 1969 ). For example, superordinate concepts ( vegetable ) are assumed to reside at a higher level than basic-level concepts ( carrot ). A feature-based attractor network with a single layer of semantic features developed representations of both basic-level and superordinate concepts. No hierarchical structure was built into the network. In Experiment and Simulation 1, the graded structure of categories (typicality ratings) is accounted for by the flat attractor network. Experiment and Simulation 2 show that, as with basic-level concepts, such a network predicts feature verification latencies for superordinate concepts ( vegetable ). In Experiment and Simulation 3, counterintuitive results regarding the temporal dynamics of similarity in semantic priming are explained by the model. By treating both types of concepts the same in terms of representation, learning, and computations, the model provides new insights into semantic memory.     

Social priming of dyslexia and reduction of the Stroop effect: What component of the Stroop effect is actually reduced?

Recently, Goldfarb, Aisenberg, and Henik (2011) showed that in a manual format of the Stroop task, dyslexia priming eliminates the normal magnitude of the interference-based Stroop-like findings otherwise exhibited by individuals participating in such research. Goldfarb et al. (2011) consequently concluded that the effect of word reading in a Stroop task (i.e., one automatic behavior) can be effectively controlled through an automatic instruction “do not read” (i.e., another automatic behavior). The present study further investigated these ideas by examining when and how dyslexia priming controls different processes involved in a Stroop task. To this end, the original finding was first replicated (Experiment 1) and subsequently extended to the vocal (instead of manual) response modality to examine whether previously reported eliminations of the Stroop effect persist with this response format (i.e., format producing larger Stroop effects). Since past work (e.g., Augustinova & Ferrand, 2012a; Brown, Joneleit et al., 2002; Ferrand & Augustinova, 2013) had suggested that various interventions were likely to reduce (rather than eliminate) the interference-based Stroop-like findings with vocal responses, a further aim of these experiments was to identify the component of these findings that dyslexia priming actually reduces. To this end, the effects of this intervention were examined in a more fine-grained variant of the Stroop task that distinguished between interference resulting from task-irrelevant processes involved in computing the lexical and semantic representations of the word (i.e., a written distractor to ignore) and task-relevant processes involved in the selection of a response (i.e., a color target to name) that are both involved in this task. In line with our past work (e.g., Augustinova & Ferrand, 2012a; Ferrand & Augustinova, 2013), the results of two experiments (Experiments 2 and 3) showed that in the vocal format, dyslexia priming reduces but does not eliminate the normal magnitude of the interference-based Stroop-like findings and that this reduction is solely due to the control of processes involved in the selection of a response (i.e., a color target to name) – processes that are known to be controllable in this format (Ferrand & Augustinova, 2013). Given that in this format, dyslexia priming had no effect on task-irrelevant processes involved in computing the lexical and semantic representations of a written distractor to be ignored – processes that are known to be automatic – further implications for the control of automatic processes via dyslexia priming are considered and an interpretation in terms of a unitary control mechanism for both the manual and vocal formats is proposed.     

Integrating the Automatic and the Controlled: Strategies in Semantic Priming in an Attractor Network With Latching Dynamics

Semantic priming has long been recognized to reflect, along with automatic semantic mechanisms, the contribution of controlled strategies. However, previous theories of controlled priming were mostly qualitative, lacking common grounds with modern mathematical models of automatic priming based on neural networks. Recently, we introduced a novel attractor network model of automatic semantic priming with latching dynamics. Here, we extend this work to show how the same model can also account for important findings regarding controlled processes. Assuming the rate of semantic transitions in the network can be adapted using simple reinforcement learning, we show how basic findings attributed to controlled processes in priming can be achieved, including their dependency on stimulus onset asynchrony and relatedness proportion and their unique effect on associative, category‐exemplar, mediated and backward prime‐target relations. We discuss how our mechanism relates to the classic expectancy theory and how it can be further extended in future developments of the model.     

Rethinking parafoveal processing in reading: Serial-attention models can explain semantic preview benefit and N+2 preview effects

During reading, some information about the word to the right of fixation in the parafovea is typically acquired prior to that word being fixated. Although some degree parafoveal processing is uncontroversial, its precise nature and extent are unclear. For example, can it advance up to the level of semantic processing? Additionally, can it extend across more than two spatially adjacent words? Affirmative answers to either of these questions would seemingly be problematic for serial-attention models of eye-movement control in reading, which maintain that attention is allocated to only one word at a time (see Reichle, 2011). However, in this paper we report simulation results using one such model, E-Z Reader (Reichle, Pollatsek, Fisher, & Rayner, 1998), to examine the two preceding questions. These results suggest the existence of both semantic preview and N+2 preview effects, indicating that they are not incompatible with serial-attention models. We discuss the implications of these findings for models of eye-movement control in reading and provide a new theoretical framework for conceptualizing parafoveal processing during reading and its influence on eye movement behaviour.     

Generalisation of fear and avoidance along a semantic continuum

Directly conditioned fear and avoidance readily generalises to dissimilar but conceptually related stimuli. Here, for the first time, we examined the conceptual/semantic generalisation of both fear and avoidance using real words (synonyms). Participants were first exposed to a differential fear conditioning procedure in which one word (e.g., “broth”; CS+) was followed with brief electric shock [unconditioned stimulus (US)] and another was not (e.g., “assist”; CS–). Next, an instrumental conditioning phase taught avoidance in the presence the CS+ but not the CS–. During generalisation testing, synonyms of the CS+ (e.g., “soup”; GCS+) and CS– (e.g., “help”; GCS–) were presented in the absence of shock. Conditioned fear and avoidance, measured via skin conductance responses, behavioural avoidance and US expectancy ratings, generalised to the semantically related, but not to the semantically unrelated, synonyms. Findings have implications for how natural language categories and concepts mediate the expansion of fear and avoidance repertoires in clinical contexts.     

Global and Local Feature Distinctiveness Effects in Language Acquisition

Various aspects of semantic features drive early vocabulary development, but less is known about how the global and local structure of the overall semantic feature space influences language acquisition. A feature network of English words was constructed from a large database of adult feature production norms such that edges in the network represented feature distances between words (i.e., Manhattan distances of probability distributions of features elicited for each pair of words). A word's global feature distinctiveness is measured with respect to all other words in the network and a word's local feature distinctiveness is measured relative to words in sub-networks derived from clustering analyses. This paper investigates how feature distinctiveness of individual words at local and global scales of the network influences language acquisition. Regression analyses indicate that global feature distinctiveness was associated with earlier age of acquisition ratings, and was a stronger predictor of age of acquisition than local feature distinctiveness. These results suggest that the global structure of the semantic feature network could play an important role in language acquisition, whereby globally distinctive concepts help to structure vocabulary development over the lifespan.     

Does Semantics Need Normativity? Comments on Allan Gibbard,Meaning and Normativity

In the book Gibbard proposes, first, that statements about meaning are normative statements and, second, that they can be given an expressivist treatment, along the lines of Gibbard’s preferred metaethics. In my paper, I examine the first step: The claim that meaning statements are to be construed as being normative, as involving ‘oughts’. Gibbard distinguishes two versions of the normativity of meaning thesis – a weak version, according to which every means implies an ought, and a strong version, according to which for every means, there is an ought that implies it. I argue that neither thesis withstands scrutiny. The weak thesis depends on assumptions about the notion of semantic correctness that the anti-normativist rejects, and the strong thesis does not solve the problems Gibbard wants it to solve: the problems of indeterminacy and meaning skepticism. I conclude that semantics does not need normativity.