The question of whether nonhuman animals participate in intentional communication has become central in the comparative research on animal communication. Current research has focused on the signaler displaying intentional behavior (Townsend et al. 2016) mostly by applying the features of the concept of intentional signals (Call and Tomasello 2007; Liebal et al. 2014). Here, we will show that current empirical evidence may pick out a signaler’s informative and communicative intention, and a recipient’s ability to understand “the meaning of the signal” linked to the signaler’s intentions, only if researchers adopt a Neo-Gricean definition of intentional communication that views communication as fundamentally inferential. However, adopting such an approach happens mainly for reasons of methodological access to intentional communication in animals and does not exclude calling out to non-inferential accounts of communication, such as the one developed by Millikan (2005).
Furthermore, we will stress that the commonly proposed behavioral criteria for intentional communication focus on the signaler’s behavior and therefore determine the presence of intentional behavior by the signaler. However, any criteria for intentional communication should also take into account the recipient’s response, which must also display instances of intentional behavior. There are two reasons for doing so. First, the recipient’s response is a constitutive part of communicative interaction, sign use, and the very nature of a sign (Millikan 2005). This holds for both animal and human communication and for both signals and linguistic utterances. We call this reason the constitutive reason. Second, not integrating the recipient’s responses nor attempting to link those to the signaler’s communicative behavior makes it difficult to evaluate whether recipients merely display instances of decision-making (Wheeler and Fischer 2012), independently of the signaler’s potential goal and intentional behavior and the signal’s meaning, rather than truly engage in intentional communication. To exclude these cases, we need a framework that allows determining whether signalers and recipients attend to each other’s behavior and interact in a flexible, voluntary manner, as evidenced in humans. We call this reason the methodological reason.
We will explore both reasons in more details in Section 2. However, before we do so, we recall the main lines of the Gricean account of human communication made prominent in animal communication research by Dennett’s (1983) influential paper. We will analyze its shortcomings and lay out Millikan’s non-Gricean account of intentional communication in Section 2 to justify implementing an analysis of the recipient’s behavior in animal communication interactions in Sections 3 and 4. We illustrate this analysis with a few examples from the animal communication literature.
Whether nonhuman primates communicate intentionally, that is, whether they rely on intentions, such as goals and beliefs about goals and knowledge states of the recipient, when they produce signals (Tomasello 2008; Liebal et al. 2014; Townsend et al. 2016), has attracted much attention recently in the comparative research on animal communication. This is in part due to the parallels that can be made with human communication (Grice 1957; Sperber and Wilson 1995).
Currently, this question is empirically mostly addressed by identifying so-called intentional signals. This term denotes signals intentionally (i.e. goal-directed) produced as opposed to cases where the signaler may not have voluntary control over signal production (e.g. when triggered by its emotional state). While authors have stressed different aspects of potentially intentional signals (Call and Tomasello 2007; Liebal et al. 2004; Leavens et al. 2005), they converge on two central criteria, the context and the timing of the signal behavior.
First, for a signal to be produced intentionally, the signaler must produce the signal for an audience, that is in a social context. Empirically, this implies that the signaler displays audience-directed behavior, such as checking the attentional state of the recipient, and audience-specific signals, such as producing the signal to allies but not foes.
Second, the signaler must display response-waiting after signal production, i.e. it must monitor the intended recipient’s behavior. This is because signalers are interested in whether their goal has been fulfilled or not. If not, the signaler is expected to display signs of persistence and/or elaboration (Tomasello 2008; Leavens et al. 2010; Cartmill and Byrne 2007; Russon and Andrews 2010; Genty et al. 2014).
In theoretical works on intentional communication, Grice (1957) described what it takes for a situation of triangular communication (i.e. communicating something to someone via a signal, Hurford 2007) to be successful such that signalers successfully deliver the message they intended to deliver, and recipients understand that message. The signaler has to display two intentions involved in signal production (see Moore 2016b):
and
The first intention is often labeled the informative intention: the signaler intends to inform the audience about something. To do so, they produce signal x because it serves the purpose of conveying the information via its meaning. The audience’s response here can be communicative or not. The first two features of intentional signals relied upon by comparative research very well pick out such a potential informative intention via the proposed empirical criteria: the signal is used socially, and response waiting is displayed.
The second intention of importance is often labeled the communicative intention of the signaler. This communicative intention makes it overt to the audience that the produced meaningful signal is important enough to extract because it was intentionally provided by the signaler; this is often referred to as ostensive or overtly intentional communication (Sperber and Wilson 1995; Scott-Phillips 2015a).
These two features of intentional communication are both about the signaler, but this is not sufficient for labeling a communicative situation an instance of intentional communication. Both signalers and recipients have to participate flexibly and voluntarily in a communicative situation, in order to make the communication successful (Brinck 2001; Carston 2002). Therefore, the recipients must also be able to display particular features. The so-called Receiver’s Capacity (Scott-Phillips and Kirby 2013: 428) seems especially important:
It can be isolated empirically by focusing on the recipient’s responses and evaluating how they are potentially based on inferences drawn by the recipient from the context and from the signaler’s behavior. During human communication, this can be exemplified by the direction in which the signaler gazes (Senju and Csibra 2008; Csibra 2010).
A final but more neglected aspect of intentional communication stressed by Grice (1975) is that communication can be viewed as a cooperative act. Thus, communication can only be successful if all participants attend to each other (Brinck 2001) to get the intended message transmitted correctly. Therefore, we need an additional feature to represent Gricean communication such as:
If signaler and recipient interact flexibly with each other, they will evaluate the communicative behavior of the other and determine how to respond to it to successfully attain their goals and/or communicate their intentions. On a behavioral level, this might cause a turn-taking between participants (Sacks et al. 1974; Levinson 1983; Kimbrough Oller 2000; Wilson and Wilson 2005). The flexible interaction should always be primarily caused by the signal’s meaning. Furthermore, this flexible interaction might require being aware of the voluntariness of the behavior of the other as opposed to a signaler just accidently producing a signal (Grice 1982).
As Dennett (1983) pointed out, an individual displaying such an informative and communicative intention must be capable of having mental states, i.e. thoughts about external, non-mental entities in the world, and to metarepresent. Such intentionality comes in reflexive levels. In theory, endless levels of intentionality can be displayed, although they are limited by individuals’ cognitive abilities.
In the situation of two signalers’ intentions involved in intentional communication, an individual must be capable of displaying second-order intentionality to display informative intentions, that is, the signaler wants the audience to know about something when producing the signal. Therefore, to ascribe a Gricean informative intention to an individual, the latter must be able to process mental states of the second order. In the original Gricean proposal, where S wants A to know that S has an informative intention (that is, S wants to make the informative intention overt), fourth-order intentionality may even be necessary: the signaler wants the audience to know that the signaler wants the audience to know about something by producing the signal.
Such requirement may exclude any nonhuman animal from potentially displaying intentional communication, as it is unclear to what extent other animals, including nonhuman primates, can really take into account each other’s mental states (Call and Tomasello 2007). While studies tackling this question bring promising answers (Hare et al. 2000; Crockford et al. 2012; Schel et al. 2013, Liebal et al. 2004), we believe that two additional points must be noted to avoid dismissing a priori the idea of nonhuman animals communicating intentionally because of the complex mental processing required to fulfill the Gricean proposal:
Ultimately, in nonhuman primates, instead of an informative intention, an imperative (making someone do something) intention and a communicative intention of second-order intentionality (such as ‘S wants the audience to recognize the signal as an expression of a communicative act’) is sufficient to ascribe to a signaler the two required intentions for intentional communication.
The proposal laid out here for dealing with the requirements of intentional communication on the signaler’s side is summing up a potential re-evaluation of the Gricean framework, and should therefore be labeled a Neo-Gricean approach. This approach emphasizes that proponents of an ostensive-inferential account of intentional communication, such as Grice himself (1957) or Scott-Phillips (2015b), have set the bar too high for cognitive requirements. Ultimately, if one seeks to apply the framework to animal communication, it is wiser not to start with the original Gricean proposal, but rather with a revised Neo-Gricean approach (e.g. Gomez 1994, Moore 2016b).
There are other options apart from a Gricean-inspired approach: Ruth Millikan (1984, 2005, forthcoming), for instance, proposes a perception-based account on communication. Laying out Millikan’s account in more detail here will enable us to elaborate on what we previously highlighted as the constitutive and the methodological reason with respect to identifying intentional communication in nonhuman animals. Millikan rejects Grice’s claim that in order to understand what a signaler means, we need to grasp its intended effect. She presents an alternative proposal based on the assumption that the meaning of linguistic signs is determined by their proliferation history, that is, their history of selection and reproduction in a certain domain. To track down the domain of a sign, communicators display perceptual processes rather than inferential ones. Since the meaning of a sign is determined by its proliferation history, and since the domain of a sign can be tracked by perceptual means alone, recognizing the signaler’s intentions is not required for understanding what is meant by an utterance. The signaler’s inner mental life is not relevant to a receiver who is primarily interested in obtaining information about the world.
According to Millikan, linguistic forms (such as words, syntactic forms, or tonal inflections) are handed down from one person to another because these forms function to coordinate receivers’ and signalers’ behavior. Millikan defines the term function here as the kind of effect achieved by the cooperative use of a linguistic form. If the effect is desirable and solves coordination problems, the linguistic form keeps itself in circulation. Millikan calls this kind of function the ‘stabilizing function’ (Millikan 1984, 2005), which is about equivalent to the semantic meaning of a linguistic form. Linguistic forms with stabilizing functions thus become conventions, and conventions correspond to reproducing lineages of cooperatively used forms. For example, the German word ‘Esel’ (donkey) has been successfully used often enough to refer in human interactions to one particular species of animal and keeps itself in circulation for this very reason. By becoming a convention, the word ‘Esel’ serves the function of referring to donkeys whether a signaler intends to do so or not. Unlike animal species, linguistic lineages also frequently acquire new functions without changing their physical forms. Thus, the word ‘Esel’ might also be used to refer to a stubborn and stupid person. This just creates a new lineage or branch with a stabilizing function. Such novel uses of conventional linguistic forms can be introduced by the signaler’s intentions. If the hearer understands the novel use by inferring the signaler’s intention, the novel use will then serve as a new coordinating function, either temporarily or permanently. The novel use may also be copied by other signalers, and may in time be directly understood by hearers without the need to unpack the signaler's’ intention.
For our purposes, two aspects of Millikan’s view on language used in communication are of special importance. First, the recipient does not need to infer any signalers’ intentions to understand conventional linguistic forms. The reason for this stems from the fact that the function of linguistic forms is not to point to the signaler’s mind but to point to the world (except for words that have the purpose of referring to the signaler’s mind, of course). For example, the function of the word ‘Esel’ is to refer to an animal and not to the signaler’s thoughts about this animal. As Millikan puts it, the hearer directly perceives the donkey through the use of the word ‘Esel’. If the word is used correctly (i.e. in accordance with its stabilizing function), it carries natural information about donkeys just as donkey shapes, donkey calls, donkey smells, or donkey faces carry natural information about donkeys. By processing this natural information by the various sense modalities, we perceive the object at the source of the natural information in question. In this sense, we perceive objects directly through language (Millikan forthcoming).
Millikan thus departs from Grice’s key insight that stimuli that are non-naturally meaningful point to the very intentions that triggered their production in the first place, since she thinks that non-natural (semantic) meaning is on a continuum with natural (informational) meaning. This account is especially compelling for the case of language learning in children, since children are able to learn, use, and understand linguistic forms before they are capable of getting signalers’ intentions, and before developing full-fledged mindreading abilities. In Millikan’s view, signalers’ intentions are not of primary interest for the hearer, and hearers understand utterances non-inferentially by using signals with conventional stabilizing functions.
It is only in problematic cases (such as incorrect, unexpected or novel use of words) that inferences to the signalers’ intentions play an important part in successful linguistic communication. In contexts of adult human communication, we always already presuppose that signalers intend to communicate something. In the case of primate communication, however, the more fundamental question arises whether we are dealing with intentional communication or not. Since problematic cases in human communication ask for inferential work on the recipient’s side, and since such problematic cases are cases that pose epistemological challenges (e.g. ‘What the heck is the signaler talking about?’), we submit that in the case of primate communication, it is necessary to take the recipient’s inference to signalers’ intentions as a methodological guide to answer the question of whether we are confronted with a case of intentional communication or not. This is what we label the methodological reason for adopting an inferential stance of communication.
Another major part of Millikan’s cooperative perception-based model of linguistic communication is that the recipient is a constitutive element of a communicative event. We previously outlined this as the constitutive reason for taking recipients’ behavior into account. In effect, all linguistic forms proliferate and acquire stabilizing functions by being used cooperatively by the signaler and the hearer. Cooperation is successful if the use of a linguistic form is followed by a certain reaction. For instance, when hearing the typical tonal inflection that is associated with a question, the expected reaction is to give an answer. Similarly, when a signaler uses the word ‘Esel’, the expected reaction of the hearer is to perceive a donkey or to think of a donkey (but not to represent a donkey in the signaler’s mind). The reaction of the hearer is therefore part of the linguistic form, because the linguistic form would not have acquired its coordinating function between hearers and signalers if the hearers had not reacted in a certain way often enough for the form to keep itself in circulation. A linguistic convention thus consists in a pattern that includes contributions by both the signaler and the hearer; in fact, the hearer’s contribution is as much a part of the convention as the signaler’s is (Millikan 2005).
The insight gained from Millikan leads us to investigate animal communication not by considering Gricean communicative intentions, but instead by considering feature (iv) of intentional communication, namely the flexible interaction between the participants of the communicative situation. Such an approach allows us to determine differences and similarities in degree between humans and nonhuman primates, relying on behavioral variables such as flexibility – caused by attending to each other – during communication. It also allows us to exclude cases in which the recipient may respond to a signal without knowing what it means, perceiving it as correlating with a certain event and therefore responding in a certain way whenever the signal occurs.
Flexible interaction means that intentional signal production should cause intentional behavior on the recipient’s side, and in some cases it means that recipients themselves produce intentional signals as a response to the signal produced by the signaler. Flexible reactions based on the signaler’s intentional behavior by the recipient may only be possible if the recipient is somehow aware that this signaler’s behavior was intentional, i.e. not accidental. This does not imply that the recipient must display extended metarepresentational abilities: within a Neo-Gricean approach, these potentially complex requirements of flexible interactions can be explained as laid out in Section 1. Such requirements therefore do not assume higher-order intentionality to be fundamental for intentional communication. What the recipient must display is a perceptual awareness of the fact that it is the signaler that causes their own (i.e. the signaler’s) behavior; additionally, it may also display awareness about the signaler having goals, a capacity within the range of the mental abilities of great apes.
One issue with such a flexible interaction-based method is that identifying the recipient’s potentially intentional follow-up behavior caused by the signaler’s communicative behavior is difficult. This is because behavior actively displayed by the recipient and caused by the signaler’s communicative behavior is difficult to distinguish from behavior involuntarily displayed or not linked to the signaler’s communicative behavior (Hobaiter and Byrne 2014; Liebal et al. 2014). Cartmill and Byrne (2007) offer a possible paradigm where orangutans were offered desirable and undesirable food, and displayed response flexibility to the experimenter’s behavior to recover their preferred food by either employing elaboration or persistence of signal use, depending on the situation. Thus, the signaler evaluated their level of success regarding the goal they intended to achieve by producing the signal, and modified their communicative behavior flexibly, depending on the recipient’s reactions. In communicative situations, the same flexible behavior would have to be displayed by the recipients, i.e. they evaluate the meaning of the signal produced by the signaler and their own motivation to do as signaled. Furthermore, in cases of diverging motivations between signalers and recipients, they should also evaluate their success of communicating their own intention, which implies some reflexive thinking present in signalers and recipients.
In this last section, we will run through a few variables that indicate this flexible interaction. Starting on the signaler’s side, the variables are (a) behavioral sequences depending on the recipient’s behavior, (b) persistence and/or elaboration behavior, and (c) stopping of signaling:
Switching to the recipient’s side, the variables are: (d) signal production depending on the environmental circumstances, (e) a behavioral pattern dependent on the signaler’s behavior, and (f) production of additional behavior to display diverging motivations:
For both communicative situations where signaler and recipient have compatible motivations, but especially where their motivations diverge, a simple turn-taking pattern should be observed. Turn-taking here refers to turn-taking of communicative and non-communicative behavior, leading signalers to reach their goals and recipients to understand the signaler’s message as well as successfully displaying their own intentions. The closest to a systematic approach to investigate such flexible turn-taking in nonhuman primate communication is perhaps found in the approaches by Hobaiter and Byrne (2014) and Rossano (2013). Hobaiter and Byrne (2014) applied a behavioral criterion they labeled ASO (“apparently satisfactory outcome”) to identify the meaning of gestures by looking at recipients’ reactions and signalers’ response behavior. Rossano applied conversation analysis to travel initiations in mother-infant dyads in bonobos. Conversation analysis examines how partner-directed behavior leads to mutual understanding (Rossano 2013: 165). Key to mutual understanding is the sequential order of behaviors directed to an individual by taking turns with that individual (see Sacks 1992; Schegloff 2007). Rossano’s approach shows how important it is to look at communication multimodally to find flexible interaction.
The focus on flexible interaction is not only important for primates, but for all animal species. Pika and Bugnyar (2011) found evidence for deictic showing and offering gestures in corvids. They also recorded follow-up behavior of signaler and recipient. As a response, the majority of recipients oriented themselves towards the signaler and the object, approached the signaler, and engaged in affiliate interactions with them. Future research, applying either a conversation analysis or ASO approach, could investigate whether such an interpretation is warranted and whether corvids indeed interact flexibly with each other to communicate each other’s goals.
Despite some skepticism towards a re-evaluation of the Gricean framework in the form of downgrading the involved reflexive level of intentionality, our review has shown that nonhuman primates, especially great apes, appear to display intentional behavior when applying a Neo-Gricean approach for informative and communicative intentions. However, we do not think that the recipient’s ability to read the signaler’s mind or engage in inference about the signaler’s intentions is constitutive for intentional communication. Instead, there are two other reasons why the recipient’s expected reaction to the signal is important for intentional communication. First, there is a constitutive reason for taking into account the recipient’s uptake in intentional communication, as illustrated by Millikan’s account of communication. Second, we think that the ability of the recipient to infer the signaler’s intention from external cues is a useful methodological guide to an answer to our question. This is the methodological reason for taking into account the recipient’s inferential capacities in intentional communicative interaction.
On an empirical note, the question whether nonhuman animals interact flexibly enough to claim that they participate in intentional communication requires further systematic research. In this respect, we believe the focus should be put on identifying communicative situations where both sides interact flexibly with each other: signaling should cause flexible responses depending on the recipient’s motivations, and flexible responses should cause further responses if motivations of signaler and recipient diverge. Currently, the kind of evidence we are looking at still appears scarce in non-primate research, but evidence is building up in the primate literature. A possible reason for the scarcity in non-primate animals may be that they do not interact that flexibly with each other. In this respect, a focus on flexibility may open a new avenue of research to investigate the possible uniqueness of primate communication, or question it.
1 The order of the authors reflects their decreasing contribution to the chapter.
Millikan’s (2005) Language: a biological model provides an overview of her work on communication and language. Liebal et al.’s (2014) Primate communication: A multimodal approach sums up empirical research on potentially intentional communication in nonhuman primates, emphasizing the importance of multimodality. For the authors’ views on identifying intentional communication in nonhuman animals, see Townsend et al. (2016) and Sievers and Gruber (2016).
Brinck, I. (2001) “Attention and the evolution of intentional communication,” Pragmatics & Cognition 9: 259–277.
Call, J., and Tomasello, M. editors (2007) The gestural communication of apes and monkeys, New York: Lawrence Erlbaum.
Carston, R. (2002) Thoughts and utterances: The pragmatics of explicit communication, Oxford: Blackwell.
Cartmill, E. A., and Byrne, R. W. (2007) “Orangutans modify their gestural signalling according to their audience’s comprehension,” Current Biology 17: 1345–1348.
Crockford, C., Wittig, R. M., Mundry, R., and Zuberbühler, K. (2012) “Wild chimpanzees inform ignorant group members of danger,” Current Biology 22: 142–146.
Csibra, G. (2010) “Recognizing communicative intentions in infancy,” Mind and Language 25: 141–168.
Dennett, D. (1983) “Intentional systems in cognitive ethology: The “panglossian paradigm” defended,” Behavioral and Brain Sciences 6: 343–355.
Genty, E., Clay, Z., Hobaiter, C., and Zuberbühler, K. (2014) “Multi-modal use of a socially directed call in bonobos,” PLoS ONE 9: 1–12.
Gómez, J. C. (1994) “Mutual awareness in primate communication: A Gricean approach,” In: S. T. Parker, R. M. Mitchell, and M. L. Boccia (eds.) Self-awareness in animals and humans: developmental perspectives, Cambridge: Cambridge University Press, 61–80.
Grice, P. (1957) “Meaning,” Philosophical Review 66: 377–388.
——— (1975) “Logic and conversation,” In: P. Cole and J. Morgan (eds.) Syntax and semantics. 3: Speech acts, New York: Academic Press, 41–58.
——— (1982) “Meaning revisited,” In: N. V. Smith (ed.) Mutual knowledge, New York: Academic Press, 223–243.
Gruber, T., and Zuberbühler, K. (2013) “Vocal recruitment for joint travel in wild chimpanzees,” PLoS ONE 8: 1–9.
Hare, B., Call, J., Agnetta, B., and Tomasello, M. (2000) “Chimpanzees know what conspecifics do and do not see,” Animal Behaviour 59: 771–785.
Hobaiter, C., and Byrne, R. W. (2014) “The meanings of chimpanzee gestures,” Current Biology 24: 1596–1600.
Hurford, J. R. (2007) The origins of meaning, Oxford: Oxford University Press.
Kimbrough Oller, D. (2000) The emergence of the speech capacity, New York: Lawrence Erlbaum.
King, B., and Shanker, S. (2003) “How can we know the dancer from the dance? The dynamic nature of African great ape social communication,” Anthropological Theory 3: 5–26.
Leavens, D. A., Russell, J. L., and Hopkins, W. D. (2005) “Intentionality as measured in the persistence and elaboration of communication by chimpanzees (Pan troglodytes),” Child Development 76: 291–306.
——— (2010) “Multimodal communication by captive chimpanzees (Pan troglodytes),” Animal Cognition 13: 33–40.
Levinson, S. C. (1983) Pragmatics, New York: Cambridge University Press.
Liebal, K., Call, J., and Tomasello, M. (2004) “Use of gesture sequences in chimpanzees,” American Journal of Primatology 64: 377–396.
Liebal, K., Waller, B. M., Burrows, A. M., and Slocombe, K. (2014) Primate communication: A multimodal approach, Cambridge: Cambridge University Press.
Millikan, R. (1984) Language, thought, and other biological categories, Cambridge, MA: MIT Press.
——— (2005) Language: A biological model, Oxford: Oxford University Press.
——— (forthcoming) Beyond concepts. Unicepts, language and natural information, Oxford: Oxford University Press.
Moore, R. (2016a) “Meaning and ostension in great ape gestural communication,” Animal Cognition 19: 223–231.
——— (2016b) “Gricean communication and cognitive development,” Philosophical Quarterly. Version of record online: 2 August 2016. doi: 10.1093/pq/pqw049.
Pika, S., and Bugnyar, T. (2011) “The use of referential gestures in ravens (Corvus corax) in the wild,” Nature Communications 2: 560.
Rossano, F. (2013) “Sequence organization and timing of bonobo mother-infant interactions,” Interaction Studies 14: 160–189.
Russon, A., and Andrews, K. (2010) “Orangutan pantomime: Elaborating the message,” Biology Letters 7: 627–630.
Sacks, H. (1992) “Adjacency pairs: Scope of operation,” In: G. Jefferson (ed.) Lectures on conversation: Volume II, Oxford: Blackwell, 521–532.
Sacks, H., Schegloff, E. A., and Jefferson, G. (1974) “A simplest systematics for the organization of turn-taking for conversation,” Language 50: 696–735.
Schegloff, E. A. (2007) Sequence organization in interaction: A primer in conversation analysis, Cambridge: Cambridge University Press.
Schel, A. M., Townsend, S. W., Machanda, Z., Zuberbühler, K., and Slocombe, K. E. (2013) “Chimpanzee alarm call production meets key criteria for intentionality,” PLoSONE 8: 1–11.
Scott-Phillips, T. (2015a) “Nonhuman primate communication, pragmatics and the origins of language,” Current Anthropology 56: 56–80.
——— (2015b) “Meaning in animal and human communication,” Animal Cognition 18: 801–805.
Scott-Phillips, T., and Kirby (2013) “Information, influence and inference in language evolution,” In: U. Stegmann (ed.) Animal communication theory, Cambridge: Cambridge University Press, 421–442.
Senju, A., and Csibra, G. (2008) “Gaze following in human infants depends on communicative signals,” Current Biology 18: 668–671.
Sievers, C., and Gruber, T. (2016) “Reference in human and non-human primate communication: What does it take to refer,” Animal Cognition 19: 759–768.
Sperber, D., and Wilson, D. (1995) Relevance: Communication and Cognition, second edition, Malden, MA: Blackwell.
Tomasello, M. (2008) Origins of human communication, Cambridge, MA: MIT Press.
Townsend, S. W., Koski, S., Byrne, R., Slocombe, K., Bickel, B., Braga Goncalves, I., Boeckle, M., Burkart, J., Flower, T., Gaunet, F., Glock, H., Gruber, T., Jansen, D., Liebal, K., Linke, A., Miklosi, A., Moore, R., van Schaik, C., Stoll, S., Vail, A., Waller, B., Wild, M., Zuberbühler, K., and Manser, M. (2016) “Exorcising Grice’s ghost: An empirical approach to studying intentional communication in animals,” Biological Reviews. Version of record online: 2 August 2016. doi:10.1111/brv.12289. [Epub ahead of print]
Wheeler, B. C., and Fischer, J. (2012) “Functionally referential signals: A promising paradigm whose time has passed,” Evolutionary Anthropology 21: 195–205.
Wich, S. A., and de Vries, H. (2006) “Male monkeys remember which group members have given alarm calls,” Proceedings of the Royal Society Series B 273: 735–740.
Wilson, M., and T. P. Wilson (2005) “An oscillator model of the timing of turn-taking,” Psychonomic Bulletin and Review 12: 957–968.