2.1 Features of emblems

Like other categories of gesture—and other communicative phenomena more generally—the category of emblems has fuzzy edges. It’s not possible to come up with a list of rigid inclusion criteria; membership in the category is graded rather than all or nothing, and some emblems seem more typical than others (for discussion, see Hanna 1996; Payrató & Clemente 2020). Here we review five features widely ascribed to emblems that serve to set them apart from other classes of gestures. It bears emphasizing that not all gestures that may be considered emblems exhibit these properties to the same degree, or on every occasion of use.

2.2 Relationship to other phenomena

As noted, the features of emblems are graded rather than all-or-nothing; the category has blurry boundaries. Emblems thus overlap in some of their key features with other phenomena, gestural and otherwise. Here we discuss these overlaps (summarized in Table 1). Contextualizing emblems in this way helps sharpen our understanding of how they fit into broader communicative repertoires and, ultimately, why we use them.

3.1 Brief history of emblem scholarship and terminology

Emblems have been discussed in different genres across millennia, from classical works on oration (Quintilian 1922 [orig. 95 C.E.]), to early Renaissance treatises on human language (Bulwer 1644), to studies of expression across species (Darwin 1872). One of the first in-depth treatments of the class was Andrea de Jorio’s (1832/2000) study of gestural communication in Naples, still cited for its insight. Another notable early contribution was David Efron’s (1941/1972) documentation of Italian emblems, conducted as part of his research comparing the gestures of immigrant communities in the US. Some consider Efron’s study a “before and after” moment in research on emblems (Payrató & Clemente 2020: 20). For a detailed historical overview of the study of emblems, see Payrató and Clemente (2020, especially Chapter 1).

Across key treatments, emblems have been discussed under a range of labels. These include:

• emblematic gestures (Efron 1941/1972)

• emblems (Ekman & Friesen 1969)

• independent gestures (Johnson 1979)

• autonomous gestures (Kendon 1983)

• quotable gestures (Kendon 1984)

• symbolic gestures (Ricci Bitti 1992)

Many of these terms highlight the semiotic properties of emblems (“symbolic”, “emblematic”); others highlight their relationship to speech (“independent”, “autonomous”) or the fact they are conventionalized (“quotable”). These terms were used to help in the literature search for our bibliography of emblem documentation. However, the term that has become most widely used—and the one we use throughout —is “emblem”. The success of this term is likely due to its use in McNeill’s gesture classification (1992: 38), which has proved dominant in the field.

3.2 Documenting single emblems

Studies exploring the form and function of specific emblems have been conducted in a range of linguistic and cultural contexts. We summarize research on 17 different emblems in Appendix 1. These studies usually involve detailed observations about the emblem, including examples of its use and comments about its cultural significance.

Compared to broader inventories, single emblem studies allow for more detailed documentation and thus highlight the challenges involved in understanding emblems. One key challenge is understanding their historical origins. Though emblems are often thought to persist over long timescales (Morris et al. 1979; Kendon 1981), and so seem to demand historical analysis, they are also ephemeral in the interactive record, and so resist such attempts. There is, however, a wealth of information about emblem use contained in archives of print and visual media, and while this “corpus” can sometimes be difficult to access and search, several studies have begun to do so. For example, Agwuele (2014) draws on popular film, while Freestone & Kruk & Gawne (2023) draw on newspaper archives. Some scholars have even attempted to leverage such sources to help reconstruct historical trajectories of specific emblems. Krüger (2004) draws on cultural artifacts to trace the history of the horns gesture in Southern Europe; Nelson (2017) uses classical Greek and Roman literature to challenge the received history of the taboo raised middle finger (aka the bird) as the digitus impudicus of Ancient Rome.

In offering detailed observations about the subtleties of single emblems—and in attending to details of method and interpretation—these studies show us that each conventional gesture contains multitudes. They also serve as a reminder that rigorous documentation of a community’s entire emblem repertoire is no trivial undertaking.

3.3 Documenting emblem repertoires

Despite the challenges of documenting even single emblems, there is a much larger literature that attempts to catalogue entire emblem repertoires. Here, we aggregate available work across different cultural and linguistic contexts and summarize these in a table (Appendix 2). Of the 113 works included in the table, the majority document emblems within a particular language. However, language is not necessarily the best level at which to document emblems, given they are often shared at cultural levels larger (e.g., nations) or smaller (e.g., subcultures) than linguistic groupings. We report the number of emblems and inclusion of images to give a broad indication of the scope of each study.

We took a deliberately broad and inclusive approach to building this list of resources. We included works that define emblems in ways that differ from our own approach. For example, some include conventionalized touch-based gestures (Lynn 2012), exclamatives (Williams 1998), and whistles (de Jorio 1832/2000). Three existing bibliographies assisted in compiling this appendix (Hayes 1957; Epstein & Raffi 2014; Payrató & Clemente 2020: 83–84), and we also included references found in the other texts. Finally, we searched academic research databases for common names that emblems are given (see §3.1). There are a small number of references in published works that we were unable to track down. We have made note of these in Appendix 3.

There are references that involve emblem-adjacent phenomena that this process has missed, such as domain-specific signs or gestures used in art and ritual that also have emblem use for these communities (e.g., Mudras). We acknowledge there are other references that may be out there, and intend to add these to a subsequent version of the list of resources.

An important observation to emerge from our survey is that emblems in some parts of the world are much better documented than others. Figure 2 presents a map that includes a point for each language where there is at least one repertoire study (in some cases, a point indicates a cluster of languages in an areal survey). There are 77 points on the map, with most representing research at the level of national languages (e.g., Israeli, Japanese, American English). ‘Large inventories’ are those where 30 or more emblems are documented; ‘small inventories’ have fewer than 30 emblems. In large regions of Asia, Africa, Melanesia, Polynesia, and the Americas there is no recorded research on emblem inventories; there is, for instance, not a single published emblem inventory from the New Guinea region, home to more than a thousand of the world’s living languages (Palmer 2018). Note that some of these points on the map represent linguistic/cultural areas that have received sustained attention (e.g., Italy), while others represent a small study involving emblem decoding only (e.g., 10 participants from Burma in Kanayama 1999). Many works also acknowledge their partial or preliminary nature (Olofson 1974; Agwuele 2014), so even those areas with a point are not necessarily well-documented.

Figure 2: Geographic distribution of published emblem inventories.

Comparative studies account for multiple pins, sometimes including regions that would otherwise be undocumented. Three such publications have an outsized effect in the map: Morris et al. (1979)’s study of gestures across Europe (14 pins); Meo-Zilio & Mejía’s (1983) extensive documentation across the Spanish speaking world (15 pins); and Kanayama’s (1999) survey of Asia and Europe (13 pins). Figure 3 shows the same map of surveyed studies but with these three publications removed, further underscoring the thinness of emblem documentation globally.

Figure 3: Geographic distribution of published emblem inventories, excluding three large comparative studies.

In what follows, we divide the works from our survey into three broad clusters based on methods—procedural inventories, observational inventories, and emblem dictionaries—and discuss each cluster in turn.

3.4 Approaches to emblem documentation: Limitations, cross-cutting issues, and prospects

When surveying the different approaches to emblem documentation, certain key issues regarding data collection and presentation come into focus.

Existing data collection methods have limitations. Encoding/decoding studies may miss nuances in use and are subject to distortions because they rely entirely on speakers’ metacommunicative judgments. Decoding-only studies—often used in large-scale comparative projects (e.g., Morris et al. 1979; Kanayama 1999)—allow for informative comparisons but not for the discovery of previously unstudied local emblems. Observational approaches allow for richer characterization, but do not always allow for confident generalizations about the distribution of emblems across cultural, geographic, or demographic boundaries. Gesture dictionaries are useful as panoramic overviews, but the entries are usually too brief to capture subtleties of meaning. Moreover, there is currently no consistent application of lexicographic perspectives in this genre (Grigorʹeva et al. 2001: 11). In many cases one’s specific research question will strongly dictate a particular approach. But in cases where the basic goal is emblem documentation, the use of complementary approaches is a valuable strategy (see discussion in Collett, 2004). Notable in this regard is Brookes’s (2004) exploration of emblem use in a community in South Africa, which drew on recordings of naturalistic conversation, interviews with speakers, and decoding tasks.

Alongside lexicographic and anthropological methods, techniques from the documentation of signed languages can also enrich the way emblems are documented and studied. We make use of the ID-gloss approach to naming emblems in this paper (see Johnston 2010 and footnote 4), but signed language researchers also have traditions of corpus annotation (Cormier & Fenlon 2014), ways of discussing variation in presentation of manual forms (Mesh & Hou 2018 provide an elegant example), and use of computer vision processing in management of video data (Cormier et al. 2019). All these practices might be fruitfully incorporated into emblem documentation.

With regard to the presentation of data, images can be an important way to illustrate emblem repertoires. But images—whether photographs or line drawings—come with limitations. Static images are not well-suited to conveying dynamic information about gesture production, though there are ways to augment them. Figure 4 shows different ways emblems with movement can be represented in images. We include the representation of the air quotes gesture from Figure 1, as well drawn versions that do not include the start position or motion arrows. We also include a photographic representation of these different presentational options. Editing a photograph to convey movement information is more difficult, and photographs are less abstracted in their representation of an emblem, and more specific in representing an individual (including possibly identifying information regarding age, skin tone, gender, etc.). In the modern publishing era, inclusion of video—whether embedded in the digital version of an article or included as supplementary material—is feasible and should be attempted wherever possible, particularly whenever movement features are relevant. The use of video can entail ethical or privacy concerns, of course, especially in situations where there is a small speaker population (e.g., Green & Woods & Foley 2011), but re-filming with actors, or using motion capture and digital animation provide potential solutions.

Figure 4: Four different ways of representing air quotes, both as line drawings and photographs.

A related limitation concerns how emblem inventories are catalogued. Many key research projects in this area were conducted in the mid-20^th century, with a lull in publications since. As a result, emblem inventories have not yet made much use of dynamic database methods. Such databases would potentially allow entries to be updated, would enable flexible navigation (e.g., by form or by meaning), and would be reconfigurable to allow for multiple outputs (e.g., an online interface and or a print dictionary). Possible models for improved digital methods may be found in the adjacent field of language documentation (Thieberger & Berez 2012). One example is the Glottobank project, which is building large cross-linguistic databases of lexical, grammatical, and phonetic systems for use in comparative work (List & Greenhill & Gray 2017).¹¹ Another set of resources that could help with presenting emblem inventories are the signbank dictionaries for signed languages. There are currently signbanks for ASL (Hochgesang & Crasborn & Lillo-Martin 2023), BSL, Auslan and other signed languages, as well as the Global Signbank being built at Radboud University, which currently includes Kata Kolok (Indonesia) and NGT (Dutch Sign Language) (Crasborn et al., 2018).¹² These signbanks integrate video performance of lexical signs with definitional information and links to related entries in a searchable database. There are, in short, exciting opportunities to present emblem repertoires in ways that are flexible and engaging for different audiences, and that address some of the shortcomings of static dictionaries.

By incorporating methods from adjacent subfields, there is also an opportunity for emblem research to formulate new questions. For example, variationist sociolinguistic methods could inform the study of how emblem repertoires vary between groups. This might entail conducting informational interviews with participants and more closely attending to factors such as a speaker’s age, location, and social networks, to better understand emblem isoglosses. Other promising approaches could come from corpus linguistics. This might involve making use of the emoji forms of emblems that accompany text, or leveraging improvements in video image detection to examine emblems in video corpora. Approaches from cultural evolutionary studies of the diffusion and transformation of cultural forms—whether folktales, writing systems, or artistic tropes—could also be adapted to better characterize the birth, spread, and transformation of emblems.

Regardless of which approach—or combination of approaches—is taken to emblems, methodological transparency is paramount. Unfortunately, much of the research to date has presented data in the form of lists, with little reporting of the methods used to arrive at those lists. Future publications should strive for clarity in describing exactly how the gesture data were collected, in keeping with best practices in linguistics and across the social sciences (Berez-Kroeker et al. 2018). Such transparency allows for the cumulative refinement of methods. It also paves the way for better comparative work across regions through replication of existing methods and the gradual adoption of field-wide standards.

4.1 How emblems mean

Emblems are generally said to have a standard meaning—one that is clear, precise, and easy to put into words. But, on deeper examination, emblems can mean in different and subtle ways. Some scholars thus distinguish subclasses of this category: lexical emblems, with a meaning that can be translated in a single word; holophrastic emblems, with a meaning that can be rendered into a larger phrase or complete utterance; and concept emblems, with a set of related meanings organized around a theme or semantic core (Poggi 2002; Kendon 2004). In her studies of emblems in urban South Africa, for instance, Brookes (2004) separates out lexical gestures such as money (hand in fist, thumb and index fingertips rubbed together) and cry (index and pinky extended and drawn down the face from the eyes, palm inward) from holophrastic emblems like return (index finger down, hand held out from body and rotated at wrist to indicate ‘I’m coming back’) and be quiet (index finger extended and placed against lips) and from concept emblems such as clever. Distinctions between these subclasses are not always easy to draw—and some question their utility (e.g., Payrató & Clemente 2020)—but this three-way division captures the fact that emblems are heterogeneous and may convey a narrower or wider breadth of meanings.

Of particular interest is the wider end of this meaning-breadth continuum—the subclass of concept emblems. Membership in this category is not always obvious, and some argue that it includes even narrow-seeming gestures like the thumbs up. Two general approaches have been proposed to account for how meaning is organized in such cases. A first is to consider these gestures as having both paradigmatic and syntagmatic meanings (Sherzer 1991; Brookes 2001). The paradigmatic (or basic) meaning is the one that sets the emblem apart from other emblems in the same class (e.g., the head shake from the head nod). The syntagmatic (or situational) meaning is the one the emblem takes on within specific interactional contexts. The paradigmatic meaning of the thumbs up gesture, for instance, may be translated as “good,” but in context the gesture may be used to answer a yes-no question or to request permission; the paradigmatic meaning of clever is the notion of “seeing” (the handshape is suggestive of two eyes emitting rays of vision; Figure 1), but in context it may be used to assert that someone is streetwise, to admonish an observer to watch out, or to greet a passer-by. In the case of clever, these different uses entail slightly different hand movements and orientations, prompting the question of whether concept emblems typically involve a wider range of forms than narrow-breadth emblems.

A second approach is to understand concept emblems as involving a conventionalized network of meanings. Such a network consists of a core meaning and extensions from that core, and in some cases extensions from those extensions (Cooperrider & Núñez 2012; Cooperrider & Abner & Goldin-Meadow 2018). Crucially, in this second approach, the extended meanings are not just emergent in context but are recognized as standard uses of the gesture, much like different senses of a word in the case of linguistic polysemy (for the idea of gestural polysemy, see Sherzer 1973). An example is the palm-up epistemic gesture, which has been argued to have a core meaning of “lack of knowledge”, but which is also used in questions, exclamations, and statements of obviousness, among other contexts (Cooperrider & Abner & Goldin-Meadow 2018). A useful analogy is with the semantics of spoken diminutives. Diminutives have a core meaning of ‘small’ or ‘child’ but are often also used to convey endearment, precision, pragmatic hedging, and other meanings (Jurafsky 1996).

In Figure 5 we illustrate these two frameworks for understanding emblematic meanings—the “syntagmatic-paradigmatic” framework and the “core-and-extensions” framework—using previously discussed examples: the palm-up epistemic (discussed by Cooperrider & Abner & Goldin-Meadow 2018) and the clever gesture (discussed by Brookes 2001). The key difference between the frameworks is that, in the core-and-extensions framework, the non-central meanings are understood to be conventionalized uses of gesture (represented by solid arrows and borders); in the syntagmatic-paradigmatic framework, the non-central meanings are understood to be emergent in context (represented by dashed arrows and borders). Note that only a selection of extended/paradigmatic meanings are represented here. Note also that both these gestures have uses that are more emblem-like (e.g., used without speech), and uses that are less emblem-like.

Figure 5: An illustration of two emblems (top row: palm-up epistemic; bottom row: clever) seen through the lenses of the “core-and-extensions” and “syntagmatic-paradigmatic” frameworks.

It is possible, of course, that both approaches are partially correct: Some extended meanings may be stored as conventional uses of an emblem, whereas others may be innovated and interpreted ad hoc. This may seem like a relatively abstract question about emblematic meaning, but it bears on broader questions about, for instance, how emblems are processed in the moment, how they are learned by children, and how they are transmitted from community to community. The very existence of the narrow-wide continuum also raises questions about how emblematic meaning changes over historical timescales. Do emblems tend to accumulate meanings and uses over time? Do emblems for certain types of concepts tend to branch out, while others tend to remain narrow? And so on.

4.2 How emblems are learned

Children begin to acquire conventionalized gestures at a young age. Some of the very first recognizable, stable communicative acts that infants produce are well-known emblems like the head shake or wave goodbye. These tend to appear around the child’s first birthday, a bit before the better studied milestone of pointing (Frank et al. 2021). Around the same time, children also begin to produce “gestural names” (Table 1). There is some evidence, in fact, that children are quicker to acquire gestural symbols of this kind than they are to acquire spoken words (Goodwyn & Acredolo 1993). Children slowly master the emblem inventory of their community as they get older. In one study, 4-year-olds showed better emblem mastery than 3-year-olds—though not yet adult-like mastery—and were better at decoding than encoding at both ages (Kumin & Lazar 1974).

Beyond these basic facts, a number of questions remain about the acquisition of emblems. Do children reliably learn certain emblems before others? If so, why? A plausible hypothesis is that children first deploy those emblems for which communicative need is the greatest. This could account, for instance, for the fact that children often start producing the head shake before the head nod (Kettner & Carpendale 2013; Frank et al. 2021). A further issue, related to the meaning-breadth continuum, is whether children understand or produce certain subclasses of emblems before others. Zooming in on concept emblems, we might also ask whether children use the core meaning of such gestures before the extended meanings.

Children integrating emblems into their communicative repertoires involves a vertical transmission process, with conventional forms passing from one generation to another. This kind of transmission has possible parallels in how emblems are learned in cases of community contact, a horizontal transmission process. It has been widely noted that emerging signed languages incorporate and adapt the gestures used by the surrounding speaking community (Haviland 2015; Mesh & Hou 2018; Le Guen & Petatillo & Kinil Canché 2020). Less noted is the fact that the gestures selected for incorporation are invariably emblems (or, in some cases, recurrent gestures or gestural practices). Mesh & Hou (2018), for instance, show that signers in San Juan Quiahije, Mexico have integrated a set of conventional negation gestures into their signing—the wag (index or all five fingers extended, palm outward, lateral movement from the elbow), twist (rotating forearm with all fingers splayed or just the thumb and pinky extended), and palm-down (palm down hand rapidly moved horizontally away from speaker), among others—while also using some of these in new ways. For example, the twist is widely used by speakers to convey absence or non-existence, but signers also use it to express denial. Another case of horizontal transmission occurs when communities of hearing people who lack a shared spoken language come into contact. Such encounters are rife with gesture (Hewes 1974), and thus potentially ripe for emblem transmission. It may be through such encounters—as well as through sustained interactions along language borders—that certain emblems spread over vast areas. (As discussed below, caution is needed, however: certain gestures may be present in large areas because they have been repeatedly, independently innovated.)

Better studied are cases of emblem acquisition in second-language learning contexts. To date, this has mostly been studied in the classroom. One study found that explicit teaching of emblems appears to be more effective than implicit (Belío-Apaolaza & Hernández Muñoz 2021). In a large survey, French teachers reported they were aware of the potential for use of emblems in cross-cultural communication but lacked resources for formal instruction (Tellier & Cadet 2013). This suggests emblem documentation has the potential to be adapted as a classroom resource (Adams 1987).

Key questions for further research concern the dynamics of these various emblem transmission processes and how they compare to each other: Are there differences in which emblems get picked up, in which meanings are likely to get picked up, and in how emblems are likely to change through transmission?

4.3 Why do we have emblems?

Perhaps the most fundamental question we can ask about emblems is: Why do we have them at all? What functions do they serve? What communicative niche do they fill? The question becomes particularly puzzling when we consider that, purportedly, many emblems can be given a direct verbal translation. Why not simply use the verbal equivalent?

Part of the answer likely lies in communicative utility of the visual-gestural channel (Kendon 1981: 142–144). Gestures are particularly useful for conveying messages when silence is needed or when environmental noise interferes with speaking. When such situations repeatedly arise—as in hunting (Hindley, 2014) or certain work environments (e.g., sawmills; Meissner, Philpott & Philpott 1975)—gestural conventions spontaneously emerge. In other cases, taboos on speech (Brookes 2014)—as during mourning periods (Kendon 1988) or in certain religious orders (Umiker-Sebeok & Sebeok 1987)—pave the way for similar conventions. Such “gesture codes” are not typically described as consisting of emblems—the term “emblem” is usually reserved for everyday, community-wide gestures—but they provide clues about the contexts in which emblems are likely to spring up. For instance, emblems used to request the check at restaurants are specialized for noisy contexts in which speech would be ineffective. A similar explanation might serve for waves and beckoning gestures, which convey messages at a distance or when speech would be disruptive.

Functional considerations like these may shed light, not just on why emblems exist, but also on why emblems seem to cluster around particular uses. Kendon (1981) found that most emblems are geared toward three broad purposes: interpersonal control, announcement of one’s state, and the evaluation of a third party. Insult emblems, for instance, seem especially widely attested. On this account, what this class of gesture is not particularly geared for is conveying specific lexical concepts. That is, communities rarely develop emblems that encode basic objects (e.g., ‘tree’ or ‘car’), properties (to our knowledge, there are no color emblems attested) or events (e.g., ‘jump’ or ‘read’). Where there are examples of emblems that encode such lexical concepts, they often carry a richer interactional meaning. The Japanese emblem money (formed with the same handshape as in ok) is used to request payment; hungry (rubbing the belly with an open hand) can be described as conveying a property, but is mostly used in interaction to request or offer food. The strength of this observed clustering toward certain functions remains to be investigated more broadly, with several studies suggesting that, contra Kendon’s claim, some communities have a relatively rich store of narrow lexical emblems (Sparhawk 1978; Brookes 2004; Le Guen & Petatillo & Kinil Canché 2020). But should the functional clustering in emblems prove robust, a natural further question is what drives it. The fact that so many emblems are used for interpersonal control, and so few are used for object or adjective reference, makes sense when we think of the messages people are likely to send from a distance or in noisy circumstances. The need to insult someone from afar or in a dense crowd likely arises more often than the need to describe the shape or colour of some entity.

Functional considerations may also shed light on why some communities have richer emblem repertoires than others. Kendon (2004) offered such an account of the expansive emblem repertoire reported in southern Italy. He argues that the density and din of Naples, as well as the cultural practice of communicating from balcony to balcony and balcony to street, favoured the development of rich gestural conventions. (Beyond physical aspects of the urban setting, Kendon also noted that Neapolitan culture involves a certain theatricality that could drive emblem use, a need to show off “what sort of ‘character’ one aspires to play in the drama of everyday life” [Kendon 2004: 354].)

To compare speech and gesture purely in terms of their communicative efficacy would be reductive, however. Other factors no doubt enter in. Another proposal is that people are sometimes motivated to use emblems in part because of a belief—however vague and subconscious—in “word magic” (McNeill 1992: 65; for classic work on “word magic,” see Tambiah 1968). On this account, spoken words are understood to have a certain potency and to leave an indelible trace; gestures do not have these same qualities, or, again, at least not to the same degree. Anecdotal evidence for this folk belief has already been noted in the case of flight attendants requiring verbal acknowledgment. As further evidence, consider that communities around the world have long developed speech taboos of various sorts (and, as mentioned, these taboos can often be skirted by substituting gestures). When people choose a gestural insult over a verbal one, they may be doing so in part to evade the worst consequences of putting such an insult “on record.” There could also be a sense in which emblems, by virtue of being visible and less frequent than their spoken counterparts, are seen as more expressive, vivid, or playful (for a related point, see Kendon 2004: 351). Viewed from this perspective, emblems find common cause with ideophones, interjections, clicks, whistles, emoji, and other “outsider” semiotic resources. All these are thought to lie beyond the core of language but make utterances—and, by extension, the person uttering them—seem lively and colourful.