Close
Communicating for knowledge: pre-verbal infants actively and selectively seek information from others in epistemic uncertainty
Marina Bazhydai
Infants are curious learners, actively probing both non-social and social environment for information. As they navigate the world, infants may encounter epistemic uncertainty, such as when insufficient information is provided or prior knowledge is not supported by others’ testimony. Pre-verbal infants’ active communicative responses to epistemic uncertainty help delineate core cognitive competencies underlying epistemic development. In two studies, we show that eleven- and twelve-month-old infants initiate active, developmentally available communicative responses to epistemic violation of expectation events and selectively seek unavailable information from more knowledgeable social partners in situations of referential uncertainty. These studies support the proposition that infants in their first year show sensitivity to knowledge distribution among social partners, exhibit epistemic vigilance in social contexts, and actively communicate with interlocutors when they need epistemic input.
Children’s interpretation of ambiguous pronouns based on prior discourse
Manuel Bohn, Khuyen Nha Le, Benjamin Peloquin, Bahar Koymen and Michael Frank
Language learning and use is a fundamentally social behavior – when children hear ambiguous or novel words they can rely on social information to infer meaning (e.g.Tomasello, 2008; E. V. Clark, 2009). As discourse unfolds, interlocutors build up common ground - a set of shared knowledge and beliefs - that serves as a background against which new utterances are interpreted (H. H. Clark, 1996). To use common ground in this way, children not only have to pay attention to what is said earlier but also with whom they have had this conversation and share this common ground. Children as young as 18 months interpret an ambiguous pronoun as referring to a previously mentioned object - even when it was absent from the scene (Ganea & Saylor 2007; Lidz, Waxman & Freedman, 2003). Children can also rely on prior discourse to learn new words. For instance, when a speaker first states that they were hungry and then asks for a wug, 2-year-olds map the novel word onto an edible object (Sullivan and Barner, 2016).
Perceiving commitments: When we both know that you are counting on me
Francesca Bonalumi, John Michael & Christophe Heintz
Can commitments be generated without promises, commissive speech acts or gestures that are conventionally interpreted as such? While we remain neutral with respect to the normative answer to this question, we hypothesized that people believe that commitments are in place when one agent has led a recipient to rely on her to do something, even without a commissive speech act or any action conventionalized as such, and this is mutual knowledge. To probe this, we presented participants with online vignettes describing everyday situations in which a recipient's expectations were frustrated by one's behavior. Our results show that moral judgments differed significantly according to whether the recipient's reliance was mutually known, irrespective of whether this was verbally acknowledged.
Establishing a common locus of attention in mother-infant-dyads: British and Ugandan infants’ early responding to joint attention skills in response to naturalistic maternal attention directing signals
Ms. Joanna Buryn-Weitzel , Ms. Maisie Thurman , Dr. Ed Donnellan , Dr. Kirsty Graham , Maggie Hoffman , Eve Holden , Charlotte Knapper , Nicole Lahiff , Sophie Marshall , Gideon Salter , Claudia Wilke , Prof. Katie Slocombe
The ability to follow another person's gaze or pointing gesture is an important skill for young infants to acquire, as it allows infants and caregivers to establish a common locus of attention: a crucial part of joint attention. Our current understanding of point or gaze following (usually referred to as 'responding to joint attention; R-JA') is largely based on studies conducted in controlled laboratory settings with infants from WEIRD populations (e.g., Morissette et al., 1995; Mundy et al., 2007). These controlled set-ups - where an adult is typically instructed to try to direct an infant's attention to an object by giving a strict predetermined sequence of cues - do not necessarily reflect the way that adults naturally direct the attention of young infants in their everyday life though. The present study therefore aimed to examine the R-JA abilities of infants in two different cultures using a more naturalistic set-up where mothers were asked to direct their infants’ attention to objects using whichever strategies they would normally use.
What is common knowledge, and what (else) is it good for?
Malinda Carpenter
There is much developmental research on joint attention, common ground, and common knowledge that shows that even infants can share attention and knowledge with others. How do they do this – what does it mean to share attention and knowledge? I present a recent theoretical framework (Siposova & Carpenter, 2019) that distinguishes different levels of social attention and knowledge and highlights the key role of second-personal interactions and communication. I also present a series of studies showing some of the many benefits of joint attention, common ground, and common knowledge (and ‘common ignorance’), not just for effective communication and coordination, but also for increasing prosocial behaviour, and for promoting joint commitments.
Common Knowledge Promotes Cooperation in the Threshold Public Goods Game by Reducing Uncertainty
Paul Deutchman, Dorsa Amir, Matthew Jordan and Katherine McAuliffe
Recent work suggests that an important cognitive mechanism promoting coordination is common knowledge—a heuristic for representing recursive mental states. Yet, we know little about how common knowledge promotes coordination. We propose that common knowledge increases coordination by reducing uncertainty about others’ cooperative behavior. We examine how common knowledge increases cooperation in the context of a threshold public goods game, a public good game in which a minimum level of contribution—a threshold—is required. Across two preregistered studies (N =4,111), we explored how varying (1) the information participants had regarding what their group members knew about the threshold and (2) the threshold level affected contributions. We found that participants were more likely to contribute to the public good when there was common knowledge of the threshold than private knowledge. Using structural equation modeling, we found that the predicted number of group members contributing to the public good and certainty about the predicted number of contributors mediated the effect of information condition on contributions. Our results suggest that common knowledge of the threshold increases public good contributions by reducing uncertainty around other people’s cooperative behavior. These findings point to the influential role of common knowledge in helping to solve large-scale cooperation problems.
Common ground as a normative condition
Bart Geurts
Since the 1960s, common ground and kindred notions have been held to exhibit a characteristic recursive structure, along the following lines: p is common ground between individuals a and b iff a Vs that p, b Vs that p, a Vs that b Vs that p, b Vs that a Vs that p, and so on; where V is taken to be a mental-state verb like “believe” or “know”. This view has been taken to entail that, in order to have common ground, a and b must both have infinitely many V-states, which, at least on the face of it, is absurd. I accept this criticism and argue that it suggests an alternative view, namely, that common ground is a normative structure rather than a structure of mental states. As a matter of fact, this was David Lewis’s view in Convention (1969), but whereas Lewis read V as “have reason to believe”, I propose to read it as “be committed to”.
Speaker adjustments to addressees during language production
Myrto Grigoroglou and Anna Papafragou
On traditional views of speech planning, speakers design utterances to match their listeners’ particular informational needs (e.g., by telling listeners things they cannot see but need to know).1 However, not all choices in production are made for a specific listener.2,3 For instance, adults mention atypical more often than typical instruments in stories (cf. stab someone with an ice pick vs. a knife) because atypical components are highly unpredictable for any ‘generic’ comprehender.2 Developmental research has focused on children’s ‘particular’ adjustments (mostly to a listener’s visual perspective) and has produced mixed findings.4,5 Here we revisit children’s adjustments in production and probe a wider array of factors inspired by the literature on adults’ speech planning.2,6,7 Focusing on mention of instruments, an optional VP constituent, we probe effects of both generic (typicality of instruments) and particular factors (listener’s visual access, conversational goals) to instrument encoding during production.
Joint Attention and Communication
Rory Harder
Joint attention occurs when two (or more) people attend together to some object. Infants begin engaging in it around their first birthday and it serves as a crucial milieu for the development of social cognition. Hence, it must be simple enough to allow more complex theory of mind abilities to develop on its basis (León 2021), yet rich enough to foster their growth. Similarly, it has a rich epistemic component---“Everything is in the open, nothing is hidden” (Peacocke 2005, p. 298)---despite being a predominantly perceptual interaction. Because of this Janus-faced nature, it has become a topic of interest among psychologists and philosophers over the past several decades. I begin by introducing the communicative conception of joint attention (Carpenter & Liebal 2011, Eilan 2015, Siposova & Carpenter 2019, León 2021), according to which it holds partly in virtue of communication about the relevant object. Crucially, the communicative conception suggests that a suitably subtle notion of communication may capture how joint attention is at once simple and complex. Identifying this form of communication takes us beyond Gricean orthodoxy, since it must not rely upon sophisticated communicative intentions, yet be a genuine communicative action and bring about the characteristic openness of joint attention. So, I conclude by outlining the needed notion of communication by appeal to our deep-seated social motivation and its phenomenology.
Cognitive mechanisms for sentence processing: From common ground to multiple perspectives
Daphna Heller
Dominant accounts of language use posit a central role for a representation of “common ground” which encodes the mutual knowledge of the conversational partners. In this talk, I argue that, while mutual knowledge is clearly relevant to communication, a representation of “common ground” cannot be the cognitive foundation of language use, as it fails to capture a range of basic linguistic and conversational phenomena. Instead, I argue that the cognitive mechanisms that support language use involve separate, distinct representations of the self and other.
The first part of the talk focuses on reference, a topic that has received much attention in the literature on perspective-taking. I present a reference production experiment which made use of visually-misleading objects, whose function is not consistent with their appearance (e.g., a crayon shaped to look like a Lego). The referential patterns produced in these contexts cannot be accounted for by appealing to “common ground”, and instead point to the simultaneous integration of both the speaker’s perspective and the listener’s perspective.
The second part of the talk will take a more macro view of language use. Based on additional empirical evidence from questions as well as memory for conversation, I argue for a cognitive architecture where representations of self and other are tracked and updated, and are also compared regularly during conversation. This theoretical shift affords expanding our empirical landscape to cover linguistic and other conversational phenomena that encode asymmetries between conversational partners.
Preschoolers’ sensitivity to knowledge differences about object existence and object identity in real-time language processing
Narae Ju, Elizabeth Morin-Lessard, Craig Chambers and Susan Graham
Successful communication requires children to both appreciate another’s perspectives and integrate that perspective information rapidly. To date, most experimental studies that have examined children’s real-time inferences in perspective-taking contexts have used an experimental paradigm involving information that is visually shared or not shared between the child and a speaker. These studies create a discrepancy between the speaker and listener's perspectives using a physical setup where, in the critical test condition, the child listener can see objects hidden from the speaker by opaque panels. Evidence from this paradigm has shown that preschoolers can integrate another’s perspective information in the early moments of language processing (Khu, Chambers, & Graham, 2019; Nadig & Sedivy, 2002). It is important to note, however, that measures of listeners' sensitivity to visual perspective may not capture the many other ways in which information about another person's knowledge state is acquired in real-world communicative interactions (e.g., via other sensory information, verbal testimony, inference, etc.). As a result, there are outstanding questions regarding how successfully children mentalize others’ perspectives in situations involving greater representational complexity (which are part of everyday communication), and how this affects the efficiency with which children apply perspective information during real-time language processing.
The early emergence of belief justification in young children
lldikó Király, Réka Schvajda, Mikolaj Hernik, Pierre Jacob, Dan Sperber and Gergely György
The aim of our study was to explore the early emergence of reason-giving ability in interactive situations. In our study we tested whether children adjust their communicative actions in a false belief situation in order to justify their own behavior, when the task at hand allows them to take into account the desires of another person. Our goal was to explore this process at an age when verbal communication is very limited (18 months), and compare their pattern of behavior with a group of children who are already competent users of language (42-48 months).During the procedure, the child and an experimenter (E) sat behind two containers with snacks (apples and carrots), which were occluded from the protagonist (P). P stated which food she preferred, E shared a piece with her and P left the room. In P’s absence, E ate some snacks, which resulted in only one piece remaining: either from P’s preferred (control condition) or non-preferred kind (experimental condition). Upon returning, P requested a snack, and in case the child gave one, she prompted them to justify their behavior. We recorded whether the participant shared any food with P, as well as participant’s communicative behavior.
Our predictions were that older children would provide the snack in the control condition, but in the experimental condition they would be more likely to decline it. In addition, we predicted that they justify their behaviour by referring to the absence of the favoured snack kind. This justification serves the purpose of updating the common ground of the interaction as well. In addition, we expected toddlers to give the remaining snack in both conditions. However, we hypothesized that in the experimental condition specifically, they will be prone to justify their action nonverbally by revealing the preferred food’s empty container more frequently, than that of the non-preferred food. So, their tendency to update the belief of the partner would not merely reflect information giving. In line with our predictions, preliminary analyses indicate that older children interpreted the two conditions differently: they declined giving any food to the protagonist in the experimental condition more often, than in the control condition (N=39; Fisher’s exact test: p=.013). Moreover, they showed sensitivity to the different epistemic states of the protagonist generated in the two conditions and tended to justify their behavior spontaneously more often in the experimental condition compared to the control condition (Fisher’s exact test: p=.001). The testings are still in progress for the 18-month-old participants, results will be reported after the closure of the data collection. In conclusion, our results indicate that children - at the age of 3 - are able to identify situations during which justifying their behavior is necessary, and spontaneously share relevant information if needed.
Tracking disruptions of common ground for belief attribution and update
Ágnes M. Kovács, Frances Buttelmann, Barbara Pomiechowska & Dóra Fogd
Social interactions require not only representing conspecifics mental states but also a continuous and flexible updating of the attributed mental content, whenever, for instance, common ground is directly disrupted (e.g. we observe that the other lost visual access to a scene we were witnessing together and a new event takes place) or indirect information indicates that our assumptions regarding what the other sees, knows or believes might be outdated or wrong. In two lines of studies we investigate these two kinds of disruptions of common ground and the computations they yield to; the first based on direct evidence that an agent lost visual access to a scene that has changed and the second based on an inference that follows the unexpected actions of the other agent that constitute a crucial signal that there is a need for belief revision.
In the first set of studies we asked what could be the potential triggering cues of belief attribution processes in a typical false belief task, and whether these processes are recruited online and spontaneously, as the events unfold. We measured the neural correlates of belief processing at different time-points using EEG in adults and found specific signatures time-locked to the first change to which the agent lost visual access (an event happening while common ground was disrupted). This data suggests that specific belief-tracking processes are triggered online at this time-point in both implicit and explicit false belief tasks.
In the second line we investigate whether people update other agents’ mental states spontaneously upon encountering an unexpected behaviour. We developed an anticipatory looking eye-tracking paradigm, in which participants had to react to a ‘partner’ who first selected a picture based on the instructions of a ‘director’ then categorized it based on its colour. We used two ambiguous (greenish/blueish) and two non-ambiguous colours (green/blue). After a practice session, the partner started to systematically miscategorize one of the two ambiguous colours. Evidence suggests that roughly half of the participants updated the partner’s mental state spontaneously, despite the fact that the task did not require them to do so (as they had a collateral task, they simply had to click on the picture that lit up after the partner selected it). They started to correctly anticipate the partner’s mistaken actions after a few observations and were able to categorize coloured items according to the partner’s ‘new’ perspective in an explicit perspective-taking task later on. Their own perspective, however, continued to exert a surprisingly strong influence on their responses, limiting their ability to spontaneously use the acquired information.
Performative and Informative Update in Assertions
Manfred Krifka
The classical account of assertion by Stalnaker (1978, 2002; Clapp 2019) implements COMMON GROUND as a context set c, a set of world-time indices that represents the shared assumptions of the participants in a conversation, and ASSERTION of a proposition φ by a speaker as having the intended effect of restricting c to those indices at which φ is true. Stalnaker makes it clear that this cannot be taken as a DEFINITION of an assertion – the update can be achieved in other ways, the speaker may have no hope or even intention to achieve it, or the addressee can reject it. Later dynamic accounts of assertions have integrated some of these additional aspects; cf. Farkas & Bruce (2010) for a mechanism of acceptance and rejection of the proposition φ using the concept of a negotiating table, and Lauer (2013) for distinct update steps, first for understanding the utterance as an expression of speaker belief and then for the update with the proposition by Gricean reasoning.
Publicity and Precursive Faith
Harvey Lederman
I give examples to introduce a notion of a proposition being public for some people, and consider the prospects of some different theories of this idea. I discuss what I call "the matching aim'' for publicity, and focus on two styles of theory which can accommodate this aim: "new attitude" and "new content" theories. I present a general puzzle for theories of publicity, based on the matching aim. I discuss possible solutions to the puzzle, and its implications for theories of publicity more generally.
Intersentential Anaphora and Situation Pragmatics
Lukas Lewerentz
Two important motivations for dynamic semantics have been intersentential anaphora, as in (1), and donkey anaphora, as in (2):
- There’s a cat. The cat is hungry.
- Every child who meets a donkey pets it.
Situation semanticists have provided non-dynamic accounts of donkey anaphora, but intersentential anaphora remain a serious challenge for situation semantics.1 I suggest that situation semantics can handle intersentential phenomena by relying on a situation-based model of conversational update.
We know that we don’t know - Children’s understanding of common ignorance in a coordination game
Hao Lucy Liu, Malinda Carpenter, Juan-Carlos Gómez
Common ground is the knowledge, beliefs and suppositions shared between partners, and is an essential socio-cognitive skill for coordinating joint actions (Clark, 1996; Tomasello, 2010). Previous research has focused extensively on what partners know they know together, i.e. ‘common knowledge’. However, another important aspect of common ground has been neglected: what partners know they do not know together, or ‘common ignorance’. Is common ignorance more difficult, or equivalently difficult compared to common knowledge? At what age can children use common ignorance to coordinate with others? What about situations in which partners have shared asymmetrical knowledge states (e.g. we know that I know X but you do not)? A new coordination game was designed to answer these questions.
Without communicating or seeing each other’s decisions, 4- to 8-year-old children (n=120; 24 at each age) played a coordination game in which they needed to make the same choice as their partner to retrieve a reward. To retrieve it, at least one of them needed to know a secret code to access the reward. The knowledge/ignorance of both partners was manipulated by openly showing one, both or neither partner the secret code, which led to four different conditions: common knowledge (both know the code), common ignorance (both do not know the code), self knowledge (child knows, partner does not) and self ignorance (child does not know, partner knows). We recorded children’s responses, together with other measurements, such as latency to respond and signs of uncertainty for each condition and age group.
Preliminary results suggest that common ignorance appears to be more difficult than common knowledge, at least at younger ages: Whereas children at all ages were at ceiling when in common knowledge with their partner, it was only by 5 years of age that children performed above chance when in common ignorance. Children also had some difficulty coordinating successfully when their partner knew the secret code but they did not (i.e. self ignorance). It was only by 6 years that they were able to do this above chance levels. We conclude that, like common knowledge, common ignorance can play an important role in coordinating with others, but may develop later. Potential reasons will be discussed in relation to the nature of ignorance and children’s strategies for solving coordination games.
Factivity, prosody, and at-issueness: Investigating the projection behavior of (non-)factives
Taylor Mahler, Marie-Catherine de Marneffe and Catherine Lai
Tonhauser et al. (2018) hypothesized in their Gradient Projection Principle (GPP) that content projects (i.e., is taken as a speaker commitment in entailment-cancelling environments) to the extent that it is does not address – or is not-at-issue relative to – the Question Under Discussion (QUD). Consistent with the GPP, at-issueness and projection have been found to be correlated for contents expressed by the complements (CCs) of factive predicates, and prosodic cues to at-issueness (via sentential information structure) also influence the projection of factive CCs (Cummins & Rohde 2015, Tonhauser 2016, Dj¨arv & Bacovcin 2020). However, prosodic cues to at-issueness appear not to have the same eect on the projection of non-factive CCs (Dj¨arv & Bacovcin 2020, Mahler et al. 2019). This dierence in behavior is consistent with the GPP if factivity is assumed to influence the at-issueness of the CC. We report on an experiment to test this
assumption. Our work goes beyond prior experimental research by investigating the relation between at-issueness and projection for non-factive CCs, and explicitly measuring effects of prosody on at-issueness.
Assertion is weak
Matthew Mandelkern and Kevin Dorst
Recent work has argued that belief is weak: the level of rational credence required for belief is relatively low. That literature has contrasted belief with assertion, arguing that the latter requires an epistemic state much stronger than (weak) belief—perhaps knowledge or even certainty. We argue that this is wrong: assertion is just as weak as belief. We first present a variety of new arguments for this, and then show that the standard arguments for stronger norms are not convincing. Finally, we sketch an alternative picture on which the fundamental norm of assertion is to say what you believe, but both belief and assertion are weak. To help make sense of this, we propose that both belief and assertion involve navigating a tradeoff between accuracy and informativity, and so it can makes sense to believe/say something you only have weak evidence for, if it is informative enough.
‘I didn’t mean to suggest anything like that!’: Plausible deniability and context re-construction
Diana Mazzarella
In their contribution to the collection ‘Mutual Knowledge’ edited by Neil Smith in 1982, Sperber and Wilson put forth a new way of thinking of the role of context in utterance interpretation. The aim of this paper is to investigate how this conception of context can shed new light on the debate in philosophy of language and linguistics on the deniability of implicit communication. Many theorists suggest that implicit communication enables a risky message to be conveyed while being deniable, disavowable, by its sender (Fricker, 2012; Pinker, 2007). This allows communicators to achieve their conversational aims while minimizing their liability and managing their reputation. Little is known, however, of the nature of denials as conversational moves, and of the factors that influence their perceived plausibility. This paper addresses the following questions: what does it mean to deny one’s intention to communicate a risky message? What makes denials more or less plausible, or credible? I argue that denials bring about a process of re-construction of the context of interpretation of the speaker’s utterance and I illustrate how considerations of cognitive utility are the key determinant for distinguishing plausible from merely possible deniability.
References
Fricker, E. (2012). Stating and insinuating. Proceedings of the Aristotelian Society Supplementary Volume LXXXVI, 61-94.
Pinker, S. (2007). The evolutionary social psychology of off-record indirect speech acts. Intercultural Pragmatics, 4(4), 437-461.
Sperber, D. & Wilson, D. (1982). Mutual knowledge and relevance in theories of comprehension. In N. Smith (Ed.), Mutual Knowledge (pp. 61-87). London: Academic Press.
Loosening repair in language logicality
Salvatore Pistoia-Reda (UPF) and Uli Sauerland (ZAS)
We focus on language logicality, i.e. the idea that language includes a deductive system. Acceptable analyticities (contradictions or tautologies) constitute problematic evidence for language logicality. In recent discussion, two main accounts have been presented to explain the available evidence ([5], [1]; [3], [4]). According to the more recent account (cf. also [2]), grammatical analyticities are accessible to the system but a pragmatic strengthening repair mechanism can apply and prevent the structures from being actually interpreted as contradictions or tautologies; this argues against the idea of a natural logic (cf. [6]). When it comes to the strengthening nature of the mechanism, however, the proposed data leaves it open whether other versions of the meaning modulation operation are required. Novel evidence we present argues that a loosening version of the repair mechanism must be available ([7]). Our observation concerns acceptable lexical contradictions that cannot be rescued if only a strengthening version of the pragmatic strategy is available; however, here we discuss whether a loosening version of the pragmatic strategy is also needed to account for some formal analyticities.
References
[1] Gennaro Chierchia. Logic in Grammar. Polarity, free choice and intervention, volume 2 of Oxford Studies in Semantics and Pragmatics. Oxford University
Press, Oxford, 2013.
[2] Gennaro Chierchia. On being trivial: Grammar vs. logic. In G. Sagi and J. Woods, editors, The Semantic Conception of Logic: Essays on Consequence,
Invariance, and Meaning. Cambridge University Press, Cambridge, forthcoming.
[3] Guillermo Del Pinal. The logicality of language: A new take on triviality, “ungrammaticality”, and logical form. Noˆus, 53(4):785–818, 2019.
[4] Guillermo Del Pinal. The logicality of language: Contextualism vs. semantic minimalism. Mind, forthcoming.
[5] Jon Gajewski. On analyticity in natural language. Manuscript, 2002. 1
[6] Salvatore Pistoia-Reda and Luca San Mauro. On logicality and natural logic. Natural Language Semantics, to appear.
[7] Salvatore Pistoia-Reda and Uli Sauerland. Analyticity and modulation. Broadening the rescale perspective on language logicality. International Review of
Pragmatics, 13(1):1–13, 2021.
Evolutionary pragmatics: A diachronic view of common ground
Paula Rubio-Fernandez
This paper explores the relationship between language and Theory of Mind (ToM), advancing the new hypothesis that pragmatic markers are a linchpin for ToM. Pragmatic markers are linguistic devices that structure discourse and mark intersubjectivity (i.e. the speaker’s assumptions about whether the listener shares their attention or knowledge). I hypothesize that pragmatic markers connect language and ToM and enable their co-development in ontogeny and co-evolution in diachrony and phylogeny through a positive feedback loop, whereby the development of one skill boosts the development of the other. To test this new account, I
propose to investigate two kinds of pragmatic markers: demonstratives (e.g., ‘this’ vs. ‘that’) and articles (e.g., ‘a’ vs. ‘the’); as well as their cultural evolution (i.e. their diachronic change through processes of learning and use).
The development of joint attention and communication from 6 to 10 months of age
Gideon Salter and Malinda Carpenter
Joint attention has played a key role in discussions of mutual knowledge and communication, from both philosophical (Campbell, 2005; Gilbert, 2007; Peacocke, 2005) and developmental (Bruner, 1986; Siposova & Carpenter, 2019; Tomasello, 2014) perspectives. However, there are still ongoing debates about both its emergence and its relation to communication. Several researchers have argued that joint attention emerges suddenly at around 9 months of age (Tomasello, 1995; Trevarthen & Hubley, 1979), whilst others argue for a more gradual emergence starting earlier in development (e.g., de Barbaro et al., 2013; Striano & Bertin, 2005). And some treat joint attention as a perceptual phenomenon, which provides a basis for mutual knowledge and enables referential communication (Campbell, 2005; Peacocke, 2005), whilst others argue that joint attention itself is a communicative phenomenon, in the sense that it consists of communicative, sharing behaviours which enable mutual knowledge (Carpenter & Liebal, 2011; Eilan, m.s.). Our aim was to progress these debates by examining the very beginnings of joint attention in development, in relation to the very beginnings of communication.
Twenty-six infants participated in a longitudinal study from 6-10 months of age, with five monthly lab visits. Mother-infant dyads took part in a free-play period and infants participated in a battery of tasks, with 3 novel tasks focused specifically on eliciting joint attention looks. These used different stimuli (unique at each session): a moving remote-controlled toy, a non-visible interesting sound (e.g., a xylophone) and silent flashing lights. We coded for communicative looks, coordinated with a facial expression and/or vocalisation, that were used to initiate joint attention, rather than gaze alternation alone. Additional tasks included assessments of other social and communicative capacities, such as gesturing, imitation and gaze and point following.
Even using this more restrictive approach, we found that over a third of infants (38%) already initiated joint attention at 6 months. There was a significant increase in the initiation of joint attention between 6 and 8 months, but no sudden increase between two consecutive months. The frequency of infants’ joint attention looks also increased gradually with age. Infants initiated joint attention significantly more often in the novel experimental tasks than in free play, suggesting that assessing infants’ ability to initiate joint attention is underestimated when assessed, as it typically is, from free play alone.
These findings provide new experimental evidence that some infants can initiate joint attention by 6 months of age. Further analyses will examine relations between joint attention and other relevant social abilities to provide new insights into the development of joint attention and communication. Together, these new findings will contribute to both philosophical and developmental debates about the ontogeny of mutual knowledge.
Grammatical encoding of shared knowledge: towards a cross-linguistic typology
Eva Schultze-Berndt, Henrik Bergqvist and Karolina Grzech
Recent cross-linguistic research has brought to light much information about grammaticalised intersubjectivity, i.e. the grammatical marking of shared/mutual knowledge, for which the term engagement has been recently coined (Landaburu 2007; Evans et al. 2018a, 2018b). The findings also shed new light on the analysis of better known cases such as the modal markers ja in German (e.g. Modicom 2012) and ju in Swedish (e.g. Bergqvist 2020). Focusing on shared knowledge of propositions rather than shared identifiability of referents (the latter signalled e.g. by definite articles), we argue that markers of shared knowledge belong to the same overall functional domain as evidentiality, and we propose a set of parameters for their analysis.
An Externalist Account of Perceptual Common Knowledge
Axel Seemann
I propose a new take on the notion of perceptual common knowledge. It involves two steps. I begin with the consideration that this kind of knowledge is made available by two or more perceivers’ joint attention to a third object. I argue that joint perceivers identify target objects by means of deictic gestures, and sometimes verbal demonstrative utterances, in an exercise of triangulation that treats the locations of co-perceivers as standpoints. Joint attention thus takes place in environments in which more than one location is a standpoint. I call this kind of environment “social space”. I then argue, in a second step, that we can explain perceptual common knowledge, in an externalist vein, as mental states of joint perceivers that are individuated by the social spaces in which they operate. I show that the resulting account has the resources to resist some recent sceptical challenges that have been raised against the possibility of common knowledge.
What Common Ground Cannot Do
Mandy Simons
In this talk, I'll challenge two of the supposed explanatory virtues of identifying context with common ground. First, I'll argue that common ground is not a good model of the information that is available to interlocutors, as has been argued by Stalnaker and by Herb Clark. Second, I'll argue that common ground is not a fully general model of what a speaker "takes for granted," and will suggest an alternative following Bratman 1992. I'll conclude by trying to tie together the two challenges by suggesting a model of linguistic exchange that emphasizes reasoning about the interlocutor's goals and assumptions over reasoning about common ground.
Pronoun interpretation in the context of dynamic action: A test of the retrieval hypothesis
Tiana V. Simovic and Craig G. Chambers
Comprehending a pronoun (she, they…) involves using linguistic and non-linguistic cues to select an intended candidate from entities in a comprehender's mental model of the discourse or situational context. Pronoun interpretation is often described as a dependency relationship involving anaphoric links to referents that have been mentioned earlier, giving rise to the notion of a “linguistic antecedent”. But what kind of information in a mental model is needed for resolving coreference? Given their status as deep anaphors [1], pronouns need not “match” linguistic antecedents with the same surface form (i.e., agreement or constituency: “I need a fork, where do you keep them?”, “Jo ran into Sue while shopping. They…”), yet the idea of retrieval is evoked in many theoretical accounts [2, 3, 4, 5]. We explored the role of the antecedent term's semantics by using novel situations where the content of this expression is no longer viable when pronoun interpretation occurs. Fig. 1 shows a visual environment where objects are located within a grid. Critically, in this context, the outcome of an instruction like “Move the house on the left to area 12” entails that the unmoved/unmentioned house is now the leftmost one. If a subsequent instruction contains a pronoun (e.g., “Now move it…”), the key point is that the antecedent expression in discourse memory no longer accurately describes the intended referent. Thus, if retrieving the antecedent term's semantics is a fundamental part of the process, some measurable processing cost should be observed relative to when the semantics are still valid, despite the intuition that the previously mentioned object is ultimately the intended referent.
Common knowledge that help is needed increases helping behavior in children
Barbora Siposova
Although there is considerable evidence that at least some helping behavior is motivated by genuine concern for others’ well-being, sometimes we also help solely out of a sense of obligation to the persons in need. Our sense of obligation to help may be particularly strong when there is common knowledge between the helper and the helpee that the helpee needs help. To test whether children’s helping behavior is affected by having common knowledge with the recipient about the recipient’s need, 6-year-olds faced a dilemma: They could either collect stickers or help an experimenter (E). We manipulated the type of knowledge (private vs. common) that children achieved about the fact that E needed their help. To examine the influence of the type of knowledge on children’s helping behavior, we used a survival analysis, which allows for the estimation of the probability of helping from two variables: whether and how quickly children helped. We found that the probability of helping when children and the experimenter had common knowledge about the experimenter’s plight (because they heard it together) was 111.5% higher than when they each had private knowledge about it (because they heard it individually)(Cox-proportional-hazards model, p=0.026). These results suggest that already in young children common knowledge can heighten the sense of obligation to help others in need.
Rethinking Common Ground
Deirdre Wilson and Dan Sperber
We outline a novel account of 'common ground' in terms of manifestness and of cognitive utility (a.k.a. relevance).
Speakers extrapolate community-level knowledge from individual linguistic encounters
Anita Tobar, Hugh Rabagliati and Holly Branigan
Speakers’ lexical choices are affected by interpersonal-level influences, like a tendency to reuse an interlocutor’s words. Here, we examined how those choices are additionally affected by community-level factors, like whether the interlocutor is from their own or another speech community (in-community vs. out-community partner), and how such interpersonal experiences contribute to the acquisition of community-level linguistic knowledge. Our three experiments tested (i) how speakers’ lexical choices varied depending on their partner’s choices and speech community, and (ii) how speakers’ extrapolation of these choices to a subsequent partner was influenced by their partners’ speech communities. In Experiment 1, Spanish participants played two sessions of an online picture-matching-and-naming task, encountering the same pictures but different confederates in each session. The first confederate was either an in-community partner (Spanish) or an out-community partner (Latin American); the second confederate was either from the same community as the first confederate or not. Participants’ referential choices in Session 1 were influenced by their partner’s choices, but not by their community. However, participants’ likelihood to subsequently maintain these choices was affected by their partners’ communities. Experiment 2 replicated this pattern in Mexicans, and Experiment 3 confirmed that these results were driven by confederates’ communities, rather than perceived linguistic status. Our results suggest that speakers encode speech community information during dialogue and store it to inform future contexts of language use, even when it has not affected their choices during that particular encounter. Thus, speakers learn community-level knowledge by extrapolating linguistic information from interpersonal-level experiences.
Effects of “We”-framing on Young Children’s Commitment, Sharing, and Helping
Jared Vasil & Michael Tomasello
Prior research suggests a “normative turn” emerges around three years of age within the context of collaborative activities. Thus far, no research has investigated how verbal framing influences children’s emerging normative psychology. The present study investigated the effects of collaborative ‘we'-framing and individualistic ‘you'-framing on preschool children’s commitment, sharing, and helping. In a sample of 96 3- and 4-year-olds, a dyadic colouring game was framed by an experimenter to the participant using either we-framing (“We will colour our papers with our markers”) or you-framing (“You will colour your paper with your marker”). Then, the participant-experimenter dyad coloured alongside one another for several minutes, identically in both conditions. Results showed a novel developmental trajectory: 4-year-olds were less likely than 3-year-olds to abandon their collaborative partner, but only following we-framing. Moreover, 3-year-olds were more likely to politely “take leave” from the task following we-framing compared to you-framing. There were no effects of framing on children’s sharing or helping behavior. These results suggest that we-framing causes young children to feel an increased sense of commitment towards their collaborative partner, though this may manifest differently across development. Taken together, these results suggest an unexplored type of linguistic context sensitivity in children’s developing normative psychology.
Lewis on common knowledge reasoning with reasons
Huub Vromen
Common knowledge plays an important role in social interactions like communication and coordination of activities. The concept of common knowledge has been criticised for being unrealistic. The basic
problem is that knowledge is something psychological, a mental state. Now that common knowledge is both indispensable and problematic, it is worth looking again at the origins of the concept. David Lewis is widely regarded as the philosopher who introduced the concept of common knowledge. His account of common knowledge differs, however, greatly from most of the later accounts. The central notions of his theory are not built on psychological states, like knowledge or belief, but on reasons to believe. Reasons to believe are normative relations between persons and propositions that cannot be reduced to mental states. Unfortunately, his account is rather informal, and the argument has some holes. In the literature, there have been several attempts to formalise his account and fill the gaps. I will argue that these accounts are missing an important aspect of Lewis’ account. Therefore, I propose a novel interpretation of Lewis’ theory that explicitly talks about reasons, using a logic inspired on justification logic.
Practical Jokes: how to do serious things with joking words
Zoe Walker
It is sometimes thought that because joking means not being serious, you cannot actually do anything serious with your words when you joke: you are, as they say, just joking. However, this is mistaken. First of all, though jokes are not serious in the sense of being sober and long-faced, they may nonetheless be serious in the sense of being sincere – you may be joking but still mean what you say. Secondly, though your explicit joking utterance may be insincere, it will nonetheless rely on serious presuppositions being in the common ground between you and your audience, which can allow you to do serious things with your joke. In this presentation, I focus on the second kind of serious joking: insincere assertions that rely on sincere presuppositions. In particular, I argue that this form of joking can allow a joker to do two serious things: to give their hearers new beliefs, and to strengthen ideologies. To argue for the former, I draw on David Lewis’s account of presupposition accommodation and various work on presupposition in fiction to show how fictional jokes can convey real-world truths to their audiences. I also comment on the possibility of using jokes to subvert and critique the ideologies on which they rely. To argue for the latter, I draw on Eric Swanson’s work on slurs to argue that similarly to slurs on his account, jokes can strengthen ideologies by emboldening both speaker and hearers in consenting to and enacting ideologies, and, unlike slurs, does so in a low-stakes way. Again, I also note that in the right context, jokes can by the same token be used to critique and subvert ideologies. Finally, I discuss some strategies of resistance for hearers who want to block these serious effects of jokes, informed by Elisabeth Camp’s discussions of flat-footed and cunning pedantry.
Low-level memory mechanism for shared history
Jessica Wang and Ian Apperly
Human beings spend much time thinking and talking about the past encounters and events. Information we once shared with others becomes common ground for future conversations. This is known as shared history (Clark & Marshall, 1981). Clark and Marshall’s classic communication theory postulates that communicators retain episodic details of the communicative episode shared with specific partners. Although it has been argued that such a high-level account is unfeasible for the burdensome processes involved (Horton & Gerrig, 2005; 2016). It is nonetheless clear that communicators do keep track of the history shared with specific partners to some degree. For instance, speakers tend to use the same referral expressions with the same audiences for past and future conversations (e.g., Brennan & Clark, 1996; Gorman et al, 2013), listeners show sensitivity to the pairing between speakers’ voices and their utterances that they have heard (Barr et al., 2014; Horton & Slaten, 2012).
It is noteworthy that much of the evidence to date has focused on the linguistic content of shared history, and that little is known about the ways in which other low-level cues could contribute and interact with the linguistic contents of an interactive episode. Furthermore, evidence suggests that listeners spontaneously attend to information in the common ground (e.g., Barr, 2008; Heller et al., 2008; Nadig & Sedivy, 2002). This highlights the possibility that attentional priority given to items in the common ground over those in the privileged ground could lead to a stronger memory record of common ground items.
We employed a widely-used computer-based referential communication task (Apperly et al., 2010; Keysar et al., 2003) followed by a surprise memory task in which participants are asked to recognise whether a given item was seen in the referential communication task. Half of the items in the memory task were old (seen in the referential communication task), the other half were new. The old items held one of three statuses: 1. Visual plus linguistic common ground: items seen by both a speaker and a participating listener and were referred to by the speaker (CGvisual+linguistic); 2. Visual common ground: items seen by both the speaker and the listener but not referred to by the speaker (CGvisual); 3. Privileged ground: items that were privileged to the listeners’ but not seen by the speaker (PG).
Analyses showed that items in the CGvisual+linguistic condition were remembered significantly better than those in the CGvisual condition. However, there was no significant difference between memory for items in the CGvisual condition versus the PG condition (Bayes factor indicated that the null hypothesis was 20 times more likely than the alternative hypothesis). This indicates that item seen by both the speaker and listener are not necessarily remembered better than those in the privileged ground. However, items that were spoken about are remembered better that those that were not. A follow-up experiment is being conducted to replicate the initial findings and to investigate the domain specificity of the memory effect. The current results suggest that information in the visual and linguistic common ground is likely prioritised in memory via a low-level memory mechanism, and that this could be a pathway to establishing shared history.
Cooperation with Police as Non-Cooperative Pragmatics
Marina Weinstein, Danielle Dionne, Nathanial Graham, Dylan Pato, and Elizabeth Coppock
Solan and Tiersma (2005) note, with puzzlement, that individuals often provide consent for property searches even when it is apparently contrary to the individual’s self-interest. Schneckloth v. Bustamonte, for example, concerned an episode in which a man consented to a search of the trunk of a car even though he knew it contained stolen checks. Does provision of consent serve as a reliable indicator of actual consent? In other words, do individuals in these situations feel forced to give consent, or is consent given freely? Cowart (2004) argues that power differentials must be taken into consideration in the evaluation of whether or not the speech act of giving consent has occurred. The present study addresses these questions through a survey probing how individuals would respond to a hypothetical search request from a police officer, depending on whether or not they are informed of their right to refuse the request. Our findings show that under risk of being charged with a crime, participants behave differently depending on whether they are informed of their right to refuse consent by a law enforcement officer. Thus a speech act of giving consent is often insincere---in a Gricean sense uncooperative---when cooperation with police is imperative.
Preschoolers’ use of emotional prosody to resolve communicative ambiguity as a function of speaker conventionality
Karolina Wieczorek, Elizabeth Morin-Lessard, Craig Chambers and Susan Graham
Consider the following utterance: “Look at my new haircut”. This utterance can convey markedly different meanings if spoken in a sad-sounding voice versus a happysounding voice (e.g., Berman et al., 2010; Morton & Trehub, 2001). As illustrated by this example, understanding another’s communicative intent often involves the integration of the words spoken with non-linguistic cues. The focus of the current study is on one particularly influential cue that speakers can use to signal their intended meaning, namely emotional prosody. Emotional prosody is a paralinguistic cue that provides information about a speaker’s emotional state or disposition as expressed through variations in pitch level, pitch contours, and rate of speech (Banse & Scherer, 1996; Frick, 1985).
The role of visual perspective-taking in pragmatic inferencing
Elspeth Wilson, Blanche Gonzalez De Linares, Ekaterina Ostachenko and Napolean Katsos
Tracking and integrating common ground, including visual perspective-taking, forms an integral part of the inferencing process in many models of pragmatic inferencing (e.g. the Epistemic Step in implicatures, Sauerland, 2004; neo-Gricean approaches more generally, e.g. Frank & Goodman, 2012). This has been extensively investigated in referential communication with adults (e.g. Heller, Grodner & Tanenhaus, 2008; Epley, Morewedge & Keysar, 2004) and, to a lesser extent, children (e.g. Nilsen & Graham, 2009); findings indicate both egocentric as well as altercentric biases that may be weighted by a variety of contextual factors (Hawkins & Goodman, 2016). The majority of work on implicatures, meanwhile, has employed experimental contexts in which informativeness and common ground align – where speaker and hearer share all relevant information that renders an utterance equally informative for both. In these situations, children become competent from 3 years with ad hoc quantity implicatures (e.g. Stiller, Frank & Goodman, 2015). Only a few studies have examined what happens when the speaker or hearer has privileged ground; these suggest adult hearers are able to integrate the speaker’s perspective and informativeness (Bergen & Grodner, 2012; Breheny, Ferguson, Katsos, 2013; Goodman & Stuhlmüller, 2013). Children, meanwhile, are sensitive to the speaker’s perspective, and able to match an otherwise under-informative utterance to a speaker who does not share all relevant information in common ground (e.g. Papafragou, Friedberg & Cohen, 2018; Kampa and Papafragou, 2019). However, no studies have investigated whether children can – like adults – not derive an implicature when critical information that would have licensed the implicature is in their privileged ground.
Humans and apes create social closeness by co-attending to the same thing in close proximity, but only humans create additional closeness through establishing common ground about this experience being shared.
Wouter Wolf
Unlike other animals, humans have a variety of ways to create social closeness with others that revolve around shared experiences. We feel socially closer to others with whom we make music, dance, play team sports and converse, especially when sharing personal information or attitudes, or by gossiping. The human ubiquity of such behaviors contrasted by the absence of similar behavior in other species raises the question what the psychological mechanisms underlying human social activities are, how these mechanisms develop in human ontogeny, and to what degree we share this psychology with other animals through common descent.
How people establish reference in three-party conversation
Xiaobei Zheng, Zhi Xia and Yuxiu Han
In conversation, in addition to using previous linguistic context, participants are also expected to make use of information that was previously shared (Clark & Marshall, 1981). In a multiparty conversation, speaker may design their speech based on the knowledge of the most ignorant partner (Yoon & Brown-Schmidt, 2018), or the combined mental states of all the partners (Yoon & Brown-Schmidt, 2019). The present study will focus on how people interpret one of their conversational partners’ perspectives in a three-party conversation. Particularly, it will explore whether and how one of the conversational partners’ perspectives influences people’s interpretation of the other partner.