Anticipating who will say what: The influence of speaker-specific memory associations on reference resolution

Some accounts of common ground assume that successful communication requires detailed consideration of others’ knowledge. In two studies, we provide evidence for an alternative account that views common ground as being mediated in part through domain-general memory mechanisms. On each trial, participants heard prerecorded instructions from one of two speakers indicating which of two displayed pictures to select. During an initial association phase, each speaker repeatedly referred to different sets of pictures. In Experiment 1, we contrasted a “between-speaker mapping” condition, in which each speaker referred to only one picture from critical item pairs (e.g., the cat drinking milk vs. the cat sitting up), and a “within-speaker mapping” condition, in which each speaker referred to both pictures within each pair, although item categories differed across speakers. On subsequent test trials, we recorded participants' eye fixations to critical displays that included both items from a category pair. Prior to the linguistic point of disambiguation, participants in the between-speaker mapping condition were more likely to fixate on the picture previously described by the current speaker, suggesting that knowledge associated with the speaker was prompting expectations for which picture would be the intended target. In Experiment 2, we used two prerecorded speakers of the same gender to strengthen the claim that the relevant implicit memory associations are speaker-specific. These results demonstrate how domain-general memory associations can be an important constraint upon language use.     

Conversation through actions and the changing of epistemic states in a game

When a person performs a certain action, it signifies that he is causing a certain event to occur. Therefore the action is conveying a certain true sentence. Playing a game is a mutual activity, namely the listener and the speaker undertake an exchange through a linguistic dialogue or communicate through action. Because of the peculiar nature of the action, the actions in games belong to an activity where the speaker speaks “true words” and the listener hears “true words.” A static game is a process through which the participants are simultaneously “speaking” and “listening”; and a dynamic game is a process where speaking and listening take place in turn. Each step of a dynamic game is a “speaking-listening” exchange. Through “listening” and “speaking,” changes in the epistemic states of the participants occur. Of course, the degree of change depends on the type of game being played. In a dynamic game, each participant proceeds through a process of induction, and thus forms new epistemic states.     

The impurity of “pure” indexicals

Within the class of indexicals, a distinction is often made between “pure” or “automatic” indexicals on one hand, and demonstratives or “discretionary” indexicals on the other. The idea is supposed to be that certain indexicals refer automatically and invariably to a particular feature of the utterance context: ‘I’ refers to the speaker, ‘now’ to the time of utterance, ‘here’ to the place of utterance, etc. Against this view, I present cases where reference shifts from the speaker, time, or place of utterance to some other object, time, or place. I consider and reject the claim that these counterexamples to the automatic indexical theory all involve non-literal uses of indexicals and argue that they cannot be explained away on the grounds that they involve conversational implicature or pretense.     

‘I’

It has traditionally been maintained that every token of ‘I’ refers to its utterer. However, certain uses of indexicals conflict with this claim, and its counterparts with respect to ‘here’ and ‘now’, suggesting that the traditional account of indexical reference should be abandoned. In this paper, I examine some proposed alternatives and the difficulties they face, before offering a new account of indexical reference. I endorse Kaplan’s view that the reference of an indexical is determined on any occasion it is used by applying its character to a particular context, arguing that the problem cases show that this is not always the context of utterance. The task facing the semantic theorist is thus to explain what fixes the reference-determining context. I consider and reject both Predelli’s suggestion that the reference-determining context is the one intended by the utterer, and Corazza et al.’s proposal that the relevant context is fixed by conventions delivered by the utterance setting. The discussion of these two accounts reveals that an adequate theory of indexical reference should allow the speaker to use indexicals in novel ways, whilst holding that what a speaker can refer to with an indexical utterance is constrained by what an audience can understand. I develop an account based around these two requirements.     

He said,she said: Episodic retrieval induces conflict adaptation in an auditory Stroop task

People respond more slowly if an irrelevant feature of a target stimulus is incompatible with the relevant feature or the correct response. Such compatibility effects are often reduced in trials following an incompatible trial, which has been taken to reflect increased cognitive control. This pattern holds only if two trials share some similarities, however, suggesting that it may be modulated by the episodic context. To look into this possibility, we had participants respond to high- or low-pitched tones by saying “high” or “low,” respectively, and ignore the simultaneously presented auditory word “high” or “low.” As expected, performance was impaired if the heard word was incompatible with the required response, and this Stroop-like effect was reduced after incompatible trials. This sequential modulation was observed, however, only if the voice in the two successive trials was the same, whereas no modulation was obtained when the speaker changed. The results suggest that sequential modulations are due to the automatic retrieval of episodic event representations that integrate stimuli, actions, and situational and task-specific control information, so that later reactivation of some elements of a given representation tends to retrieve the other elements as well.     

I Said,You Said: A Communication Exercise for Couples

The “I Said, You Said” technique leads the couple through a series of communication exercises that emphasizes the power of verbal and non-verbal cues. Initially, non-verbal cues that a couple uses to interpret each other’s spoken words are decreased to reduce the outside influences of attributed meaning. This allows the couple to focus on the clarity of the spoken message. Additional steps in this technique include restructuring speaker and listener roles, education about communication patterns, and learning how to communicate with more clarity and effectiveness, even when topics are emotionally laden. A brief vignette using the intervention follows this discussion.     

An oscillator model of the timing of turn-taking

When humans talk without conventionalized arrangements, they engage in conversation—that is, a continuous and largely nonsimultaneous exchange in which speakers take turns. Turn-taking is ubiquitous in conversation and is the normal case against which alternatives, such as interruptions, are treated as violations that warrant repair. Furthermore, turn-taking involves highly coordinated timing, including a cyclic rise and fall in the probability of initiating speech during brief silences, and involves the notable rarity, especially in two-party conversations, of two speakers’ breaking a silence at once. These phenomena, reported by conversation analysts, have been neglected by cognitive psychologists, and to date there has been no adequate cognitive explanation. Here, we propose that, during conversation, endogenous oscillators in the brains of the speaker and the listeners become mutually entrained, on the basis of the speaker’s rate of syllable production. This entrained cyclic pattern governs the potential for initiating speech at any given instant for the speaker and also for the listeners (as potential next speakers). Furthermore, the readiness functions of the listeners are counterphased with that of the speaker, minimizing the likelihood of simultaneous starts by a listener and the previous speaker. This mutual entrainment continues for a brief period when the speech stream ceases, accounting for the cyclic property of silences. This model not only captures the timing phenomena observed in the literature on conversation analysis, but also converges with findings from the literatures on phoneme timing, syllable organization, and interpersonal coordination.     

Perceptual scaling of voice identity: common dimensions for different vowels and speakers

The aims of our study were: (1) to determine if the acoustical parameters used by normal subjects to discriminate between different speakers vary when comparisons are made between pairs of two of the same or different vowels, and if they are different for male and female voices; (2) to ask whether individual voices can reasonably be represented as points in a low-dimensional perceptual space such that similarly sounding voices are located close to one another. Subjects were presented with pairs of voices from 16 male and 16 female speakers uttering the three French vowels “a”, “i” and “u” and asked to give speaker similarity judgments. Multidimensional analyses of the similarity matrices were performed separately for male and female voices and for three types of comparisons: same vowels, different vowels and overall average. The resulting dimensions were then interpreted a posteriori in terms of relevant acoustical measures. For both male and female voices, a two-dimensional perceptual space was found to be most appropriate, with axes largely corresponding to contributions of the larynx (pitch) and supra-laryngeal vocal tract (formants), mirroring the two largely independent components of source and filter in voice production. These perceptual spaces of male and female voices and their corresponding voice samples are available at: section Resources.     

The response order effect: People believe the first person who remembers an event

When groups of people remember an event, the order in which they discuss their memories is important. In three experiments, a response order effect was shown in which participants believed the first speaker to be more accurate and more confident than a subsequent speaker. Further, participants were more likely to report as their own memory what the first speaker reported than what a subsequent speaker reported. The experiments showed that the response order effect was not due to intrinsic characteristics of what the first speaker said. Even when participants chose the response order themselves and the speakers’ dialogue was counterbalanced, participants still believed that the first speaker was more accurate and confident than a subsequent speaker. Because in most situations the person who introduces a particular topic into a discussion is more accurate, people may assume that this is true, even when the response order is random.     

Identifying unreliable informants: do children excuse past inaccuracy?

In three experiments (N = 123; 148; 28), children observed a video in which two speakers offered alternative labels for unfamiliar objects. In Experiment 1, 3‐ to 5‐year‐olds endorsed the label given by a speaker who had previously labeled familiar objects accurately, rather than that given by a speaker with a history of inaccurate labeling, even when the inaccurate speaker erred only while blindfolded. In Experiments 2 and 3, 3‐ to 7‐year‐olds showed no preference for the label given by a previously inaccurate but blindfolded speaker, over that given by a second inaccurate speaker with no obvious excuse for erring. Children based their endorsements on speakers’ history of accuracy or inaccuracy irrespective of the speakers’ information access at the time, raising doubts that children made mentalistic interpretations of speakers’ inaccuracy.     

A problem for Russellian theories of belief

Russellianism is characterized as the view that ‘that’-clauses refer to Russellian propositions, familiar set-theoretic pairings of objects and properties. Two belief-reporting sentences, S and S*, possessing the same Russellian content, but differing in their intuitive truthvalue, are provided. It is argued that no Russellian explanation of the difference in apparent truthvalue is available, with the upshot that the Russellian fails to explain how a speaker who asserts S but rejects S* can be innocent of inconsistency, either in what she says or, at least, in what she implicates. Yet, while there is no semantic or pragmatic explanation of the substitution failure consistent with Russellianism, there remains the possibility of a purely psychological explanation that is, nonetheless, Russellian. This is an attractive option. It comes at a cost, however, since, in abandoning the project of providing a semantic or pragmatic explanation of anti-substitutivity intuitions, the Russellian is no longer in the business of explaining how a rational, well-informed speaker, with no incentive to mislead, can avoid inconsistency in reporting the facts as they appear.     

Linguistic communication and the semantics/pragmatics distinction

Most people working on linguistic meaning or communication assume that semantics and pragmatics are distinct domains, yet there is still little consensus on how the distinction is to be drawn. The position defended in this paper is that the semantics/pragmatics distinction holds between (context-invariant) encoded linguistic meaning and speaker meaning. Two other ‘minimalist’ positions on semantics are explored and found wanting: Kent Bach’s view that there is a narrow semantic notion of context which is responsible for providing semantic values for a small number of indexicals, and Herman Cappelen and Ernie Lepore’s view that semantics includes the provision of values for all indexicals, even though these depend on the speaker’s communicative intentions. Finally, some implications are considered for the favoured semantics/pragmatics distinction of the fact that there are linguistic elements (lexical and syntactic) which do not contribute to truth-conditional content but rather provide guidance on pragmatic inference.     

"Feminism Lite?" Feminist Identification, Speaker Appearance, and Perceptions of Feminist and Antifeminist Messengers

Drawing on a communications model of persuasion ( Hovland, Janis, & Kelley, 1953 ), this study examined the effect of target appearance on feminists' and nonfeminists' perceptions of a speaker delivering a feminist or an antifeminist message. One hundred three college women watched one of four videotaped speeches that varied by content (profeminist vs. antifeminist) and target appearance ("feminine" vs. "masculine"). Self-identified feminists responded more favorably to a profeminist message when it was presented by a feminine than a masculine-appearing speaker and expressed less feminist attitudes after viewing a masculine-appearing feminist speaker. Nonfeminists' evaluations did not vary as a function of speaker appearance or message content. Implications for the communication of feminist ideology are discussed.     

DiapixUK: task materials for the elicitation of multiple spontaneous speech dialogs

The renewed focus of attention on investigating spontaneous speech samples in speech and language research has increased the need for recordings of speech in interactive settings. The DiapixUK task is a new and extended set of picture materials based on the Diapix task by Van Engen et al. (Language and Speech, 53, 510–540, 2010), where two people are recorded while conversing to solve a ‘spot the difference’ task. The new task materials allow for multiple recordings of the same speaker pairs due to a larger set of picture pairs that have a number of tested features: equal difficulty across all 12 picture pairs, no learning effect of completing more than one picture task and balanced contributions from both speakers. The new materials also provide extra flexibility, making them useful in a wide range of research projects; they are multi-layered electronic images that can be adapted to suit different research needs. This article presents details of the development of the DiapixUK materials, along with data taken from a large corpus of spontaneous speech that are used to demonstrate its new features. Current and potential applications of the task are also discussed.     

Perelman, ad Hominem Argument, and Rhetorical Ethos

Perelman’s view of the role of persons in argument is one of the most distinctive features of his break with Cartesian assumptions about reasoning. Whereas the rationalist paradigm sought to minimize or eliminate personal considerations by dismissing them as distracting and irrelevant, Perelman insists that argumentation inevitably does and ought to place stress on the specific persons engaged in an argument and that the relationship between speaker and what is spoken is always relevant and important. In taking this position, Perelman implicitly revives the classical conception of proof by character (ethos or “ethotic” argument), but despite an extended discussion of act and person in argument, The New Rhetoric does not give much consideration to the classical concept and confuses differing approaches to it within the tradition. The result is that Perelman treats the role of the speaker in argument only by reference to abstract techniques and does not recognize the importance of examining particular cases in order to thicken understanding of how ethotic argument works in the complex, situated context of its actual use. Consequently, Perelman’s account of the role of persons in argument should be supplement by reference to case studies, and to that end, I consider ethotic argument in W.E.B. Du Bois’ famous essay “Of Mr. Booker T. Washington and Others”.

Individualism, Externalism and Idiolectical Meaning

Semantic externalism in contemporary philosophy of language typically – and often tacitly – combines two supervenience claims about idiolectical meaning (i.e., meaning in the language system of an individual speaker). The first claim is that the meaning of a word in a speaker’s idiolect may vary without any variation in her intrinsic, physical properties. The second is that the meaning of a word in a speaker’s idiolect may vary without any variation in her understanding of it. I here show that a conception of idiolectical meaning is possible that accepts the “anti-internalism” of the first claim while rejecting (what I shall refer to as) the “anti-individualism” of the second. According to this conception, externally constituted idiolectical meaning supervenes on idiolectical understanding. I begin by trying to show that it is possible to disentangle anti-internalist and anti-individualist strands of argument in Hilary Putnam’s well-known and widely influential “The Meaning of ‘Meaning’.” Having once argued that the latter strand of argument is not cogent, I then try to show that individualism (in the sense above) can be reconciled with perhaps the most plausible reconstruction of Putnam’s well-known and widely accepted “indexical” theory of natural kind terms. Integral to my defense of the possibility of an individualist externalism about idiolectical meaning are my efforts to demonstrate that, pace Putnam, there is no “division of linguistic labor” when it comes to the fixing the meanings of such terms in a speaker’s idiolect. The fact that average speakers sometimes need defer to experts shows that not reference, but only reliable recognition of what belongs in the extension of a natural kind term is a “social phenomenon.”

Wittgenstein (1958, 14).

     

Tracking the what and why of speakers’ choices: Prosodic boundaries and the length of constituents

The rational speaker hypothesis (Clifton, Carlson, & Frazier, 2002) claims that speakers are selfconsistent, employing intonation in a manner consistent with their intended message. Preceding a constituent by a prosodic boundary that is not required by the grammar often signals that this constituent is not part of the immediately preceding phrase. However, speakers tend to place prosodic boundaries before and after long constituents. The question is whether prosodic boundaries will have a larger influence on listeners’ choice of an analysis when they flank short constituents than when they flank long ones. The results of two listening experiments indicate that they do, suggesting that listeners attend not just to properties of the input signal, but also to the reasons why speakers produce those properties.