1.1 What Are Signed Languages?

Signed languages are the natural languages used by communities of deaf people around the world. Although only scientifically studied since around the mid twentieth century, these signed languages have a long history of having been observed and written about. Deaf people and signed language were mentioned in ancient Egyptian writings from the nineteenth dynasty, c.1350–1200 BCE, where it was observed that deaf people “speak” with their hands. Signed languages are mentioned in Talmudic law, and in the writings of Aristotle and Quintilian. Writing about Roman oratorical gestures, Quintilian observed that movements of the hand and even a nod may express meaning; he believed that gestures such as this are used by deaf people as a substitute for speech.

Throughout history, deaf people have gathered into groups in larger cities. We have a few historical descriptions of signed language communities, such as that of Pierre Desloges (1779 as reported in Lane & Grosjean Reference Lane and Grosjean1980, pp. 123–124). Desloges became deaf at the age of seven from smallpox. As an adult, described the Parisian deaf community, noting that:

The predominant historical view of signed languages has been that they are not language. Language, it was assumed, is spoken. Signs are produced with the hands, like gesture, and thus they were classified as holistic gestures – pantomimic images drawn with the hands lacking linguistic structure. One hundred years after Desloges wrote about the absolutely normal language abilities of deaf people, the view among those charged with educating deaf children was still that signs are merely gestures. In a speech to the Milan Conference of 1880 on the topic of methods for teaching deaf children, Giulio Tarra, the president of the conference, proclaimed (in Lane Reference Lane1984, pp. 393–394):

Linguists now know a great deal about signed languages. We have learned about their grammatical structure; their geographic diversity and distribution; the presence of large communities of signed language users such as American Sign Language (ASL), Argentine Sign Language (LSA), Brazilian Sign Language (Libras), Iranian Sign Language (ZEI), as well as much smaller communities of village languages such as San Juan Quiahije Chatino Sign Language (a village sign language used in the region of southern-central Oaxaca, Mexico), and Kata Kolok (a signed language used in a small traditional farming village in the north of Bali); and how these visually transmitted languages are perceived, acquired, and processed.

Sign linguists have demonstrated the many properties signed languages share with spoken languages, as well as some of the grammatical phenomena specific to signed languages that derive from their unique mode of production using visible hands, faces, and bodies – features such as constructed action (the use of the signer’s body to depict the actions and movements of some person or entity), the semantic and phonological use of space, simultaneous constructions (the two hands producing different content), body partitioning (different parts of the body expressing different semantic content), and more.

1.2 Signed Languages Are Linguistically Structured

The view that signed languages lack linguistic structure can be expressed in the claim that they do not exhibit duality of patterning (Pulleyblank Reference Pulleyblank1987; Pulleyblank Reference Pulleyblank2011), the feature of language in which the meaningful elements, that is, words or morphemes, are composed of a small stock of meaningless elements, that is, sounds or phonemes (Hockett Reference Hockett and William1982). In other words, the claim was that signed languages have no phonology – the very name phonology, after all, refers to “sound.”

William C. Stokoe dispelled this view in his pioneering work Sign Language Structure: An Outline of the Visual Communication Systems of the American Deaf (Stokoe Reference Stokoe2005)Footnote ¹ by demonstrating that signs do have a level of sublexical structure equivalent to the phonemes of spoken words – that is, they exhibit phonology. Stokoe called these minimal elements cheremes, the root word cher- from Homeric Greek meaning “hand.” He defined cheremes as “that set of positions, configurations, or motions which function identically [to phonemes] in the language” (Stokoe, Reference Stokoe2005, p. 33).

Stokoe argued that the phonological units of signs fall into three major classes: handshape (the configuration that the hand makes when producing the sign), location (the place where the sign is produced, for example on the chest), and movement (the motion made by the signer in producing the sign, for example a path movement away from the signer’s body). Stokoe called these three aspects of a sign, observing that, unlike the sequential ordering of phonemes in spoken languages, cheremes occur simultaneously and cannot be produced independently: it is not possible to articulate a movement without also articulating a handshape to move or a place to make the articulation. Later these units came to be called sign parameters, and as more research was conducted, more parameters were identified. Battison (Reference Battison1978) recognized a fourth parameter, orientation (the direction the palm faces when a sign is produced). We will return to Stokoe’s work and the topic of sign phonology in Section 7.1.

1.3 Signed Languages and Cognitive Linguistics

The field of sign linguistics has advanced enormously since the 1960s. Linguists working within different theoretical approaches and subfields of linguistics have contributed to our knowledge of the structure of the world’s signed languages. An early and influential book on ASL linguistics, The Signs of Language, was published by Edward Klima and Ursula Bellugi and their colleagues (Reference Klima and Bellugi1979). Not long after, a series of influential books was published that ignited the field of cognitive linguistics, including George Lakoff and Mark Johnson’s Metaphor We Live By (Reference Lakoff and Johnson1980), Lakoff’s Women Fire and Dangerous Things (Reference Lakoff1987), and Ron Langacker’s Foundations of Cognitive Grammar (Reference Langacker1987). Linguists soon began applying cognitive linguistic theory to the study of signed languages. Janzen and colleagues (Reference Janzen, O’Dea and Shaffer2001, Reference Janzen, Shaffer, Meier, Quinto and Cormier2002, Reference Janzen, Kristiansen, Achard, Dirven and de Ibáñez2006, Reference Janzen, Steinbach, Pfau and Woll2012) examined the linguistic use of space, the construal of events, and the cognitive factors at work in lexicalization and grammaticalization in signed languages. Working with Danish Sign Language, Engberg-Pedersen (Reference Engberg-Pedersen1993, Reference Engberg-Pedersen, Allwood and Gärdenfors1999, Reference Engberg-Pedersen2015) contributed to our understanding of the role of space and perspective. Shaffer (Reference Shaffer2002, Reference Shaffer2004, Reference Shaffer, Janzen, Nuyts and van der Auwera2016) applied force dynamics to the study of modality and advanced the study of evidentiality in ASL. Liddell (Reference Liddell, Emmorey and Riley1995, Reference Liddell1998, Reference Liddell and McNeill2000a, Reference Liddell2003) pioneered the application of conceptual blending theory to ASL. Expanding on this work, Dudis (Reference Dudis2004) was the first to describe the conceptual underpinnings of body partitioning and investigated its role in conceptual blends.

Metaphor, metonymy, and iconicity have long been important domains of study in cognitive linguistics. These areas were recognized and studied by sign linguists working in other theoretical approaches (Brentari Reference Diane.2007; Emmorey Reference Emmorey2014; Frishberg Reference Frishberg1975; Wilbur Reference Wilbur1987). Sign linguists working within cognitive linguistic theory have continued to explore these phenomena and their interactions in signed language lexicon and grammar across a number of languages (Emmorey Reference Emmorey2014; Meir Reference Meir2010; Nilsson Reference Nilsson2016; Perniss Reference Perniss2007; Taub Reference Taub2001; Wilcox Reference Wilcox2000; Wilcox & Wilcox Reference Wilcox and Wilcox2003; Wilcox et al. Reference Wilcox, Wilcox and Josep Jarque2003).

1.4 Goals of This Element

We have three goals in writing this Element. We have for years worked within the theory of Cognitive Grammar. Although widely influential among spoken language cognitive linguists, Cognitive Grammar has not had a large impact within the field of sign linguistics. Thus, one of our goals is to provide sign linguists with the basic tenets of Cognitive Grammar and apply it to a range of signed language grammatical phenomena. Second, we hope to show readers with a background in cognitive linguistics and perhaps even Cognitive Grammar, but who have little or no experience with signed language data, examples of the kinds of unique phenomena sign linguists confront.

Our third goal addresses the question of how we approach the study of signed linguistics. Signed languages are languages with unique properties that set them apart from spoken languages. Since the beginning of their study within the modern science of linguistics, the modus operandi of the sign linguist has often been to take theories devised to account for spoken language structure, and search for analogies in signed languages. This was even the strategy first employed by Stokoe, who turned to the structuralist linguistics of the day to show that signs have the equivalent of phonemes and phonology. We believe it is time to adopt a new strategy, one which looks at signed languages with fresh eyes.

The great advantage of Cognitive Grammar, in our opinion, is that it provides a set of conceptual tools at an appropriate level of abstraction to implement this new strategy. We acknowledge, of course, that signed languages are human languages, and connections with spoken languages will surely be found – again, at the appropriate level of abstraction. Our cognitive linguistic perspective tells us to work with the assumption that these connections will reflect the conceptual abilities that all human share, abilities that are built up from our embodied sensorimotor experience with the world.

In the following sections we will provide a background on the concepts of Cognitive Grammar and its application to the analysis of various signed language phenomena. Before beginning that journey, Section 2 offers a proposal for how to approach the study of language, and especially of signed languages, from a cognitive perspective by reviewing what we know about embodied human perceptual and motor experience. Section 3 gives a background on key concepts of Cognitive Grammar, including the cognitive models that are pervasive in the theory.

Section 4 begins the application of Cognitive Grammar to signed language grammar. We introduce several new analytic concepts, drawn from Cognitive Grammar principles but unique to signed languages and the fact that they incorporate meaningful locations in space. We show its application to such grammatical phenomena as nominal grounding; antecedent-anaphor constructions; and referent tracking, including constructions that mark conceptual overlap. One example of conceptual overlap examined in detail is role shift, in which the signer can represent a narrator and various other characters in reported dialogue and narrative.

Section 5 offers a case study of grammatical modality, suggesting a possible source of modals in gesture. Data is drawn from Brazilian Sign Language, Italian Sign Language, and the depiction of gesture in Renaissance art. Section 6 looks at facial displays, relying on a Cognitive Grammar model called the control cycle and the concepts of effective control (striving to influence what happens in the world) and epistemic control (constructing and continually updating a conception of reality). The case is made that across many signed languages, effective control is often marked by a browed furrow facial display, and epistemic control is indicated by a mouth display made by depressing or pulling down the corners of the mouth. In Section 7 we return to Stokoe and examine the impact of his work not only on phonology but also on our understanding of certain grammatical constructions. One surprising new trend has been the re-emergence of gesture, with the claim that many sign constructions are fusions of language and gesture. We close this Element with a discussion of the relation between language and gesture informed by a cognitive linguistic perspective.

3.1 Interlude: Space, Time, and Language

Language both occurs in space and time and represents space and time. As a physical activity, language takes place in space and time at several different levels of organization. At the contextual level, the speaker is situated spatially and temporally when she speaks. The act of signing or of speaking involves the movement through space and time of articulators. The neural activity which underlies speaking and signing also has a spatial and temporal dimension. Language also represents space and time: we use language to talk about objects and events which exist in space and time. Even when we talk about concepts that have no referents in the physical world, we typically conceptualize these metaphorically as objects or events (Lakoff Reference Lakoff1987; Lakoff & Johnson Reference Lakoff and Johnson1980; Lakoff & Johnson Reference Lakoff and Johnson1999; Johnson Reference Johnson1987). As we will see in more detail, it is even possible to characterize two of the basic grammatical elements of language, nouns, and verbs, as abstract schemas of space and time.

Space, time, matter, and motion are essential elements of our perceptually and experientially grounded world. Objects are located in space and time, and they move through space and time. Events and actions, although they depend on things such as people, trees, animals, and other material entities, consist of a change that unfolds through time.Footnote ⁵ In cognitive approaches such as Cognitive Grammar, the claim is that our perceptually grounded experience of the world has direct significance for grammar.

The dimensions of space and time in which language is realized have differing roles within linguistic representation. At the contextual level, for example, space and time are important aspects of language because they form the ground for the communicative act. The speaker’s or signer’s location in space finds linguistic expression as a part of the ground (here, there, behind, etc.), as does the speaker’s location in time (now, yesterday, tomorrow).

Space and time have special significance for signed languages. Not only are lexical signs produced in space; space plays a ubiquitous role in expressing grammatical relations and constructions. We will see in Section 7 that the linguistic status of locations in space is currently highly controversial. Just as for spoken languages, signed languages are expressed in time. While spoken languages often metaphorically express time in terms of space (e.g., what are the events that are ahead of us? or That was way back in 1900), signed languages express time literally in space (Engberg-Pedersen Reference Engberg-Pedersen, Allwood and Gärdenfors1999; Fischer & Gough Reference Fischer and Gough1978; Frishberg & Gough Reference Frishberg and Gough2000; Pereiro & Soneira Reference Pereiro and Fernández Soneira2004; Stokoe et al. Reference Stokoe, Casterline and Croneberg1965).

3.2 The Experiential Grounding of Language and Grammar

The view that we present is one of language and grammar as grounded in experience, embodied in perception and action; that is, it is built-up from experience rather than built-in as an innate cognitive module or self-contained ability. The perceptual events at which our embodied experience is directed consist of objects and events in the world. Things such as rocks, trees, animals, and other people are located in places; they move about (rocks roll, trees sway); they perform actions on other things (rocks tumble down hills and knock down trees). As we have noted, our fundamental drive is to make sense of this perceptual world. We try to figure out what these things are, where they are located, how they move about, what they are doing.

These are the perceptible elements in the world from which conceptual archetypes emerge. Conceptual archetypes are pervasive and experientially grounded “complex notions, intermediate in their level of abstraction, that pertain to fundamental aspects of moment-to-moment experience” (Langacker Reference Langacker, Yamaguchi, Tay and Blount2014, p. 31). A partial list of conceptual archetypes includes: a physical object, an object in a location, an object moving through space, the human body, the human face, a whole and its parts, a physical container and its contents, seeing something, holding something, handing something to someone, exerting force to effect a desired change, a face-to-face social encounter (Langacker Reference Langacker2008). Conceptual archetypes may be manifest in the mental domain, such as the experience of perceiving, of thinking, and of feeling emotions, or in the physical domain as a type of “force-dynamic” experience: muscular exertion, being subjected to forces, resisting and overcoming forces, or carrying out a physical action (Langacker Reference Langacker2000). Although the term focuses on their role in conception, conceptual archetypes consist of meanings and perceptible forms. They always come to the perceiver as a pairing of a perceptible form, and some basic meaning: this is a person, that is a tree; that person located far from me; the rock is located near to me; the animal is moving toward me. Conceptual archetypes are forms with meaning: they are archetypal symbolic structures.

Grammatical structure is grounded in conceptual archetypes. For example, one conceptual archetype involves a global setting in which mobile participants appear in certain locations, and then interact in some event in which energy is transferred from one participant to the other. This conceptual archetype forms the experiential basis of clause structure, including archetypal grammatical roles such as agent and patient, a transitive verb, and so forth.

We can think of conceptual archetypes in terms of baseline and elaboration. A baseline structure is more substantive and has some priority. A physical object in a location is a baseline conceptual archetype. An object moving through space is one type of elaboration. Further elaborations result in the global setting archetype with participants (things) in locations interacting and transferring energy.

This analysis also applies to sign phonology. Hands are substantive things with certain properties, or aspects: the hand’s configuration, location, and orientation. Elaborations can operate on each of these properties. Bending or fluttering the fingers is an elaboration of handshape, as is changing the hand’s orientation. Movement can elaborate the hand’s location. As we’ve seen, elaborations can operate on BE structures. An example of such a higher level elaboration is manner of movement, which elaborates movement, which in turn elaborates location. This analysis also suggests that the basic phonological elements of signed languages are conceptual archetypes: hand (object), location (object in a location), and movement (object moving through space). This approach accommodates facial articulators as well. The human face is a baseline element (Langacker Reference Langacker2008). Elaborations are often expressed as verbs operating on a baseline structure: we stress a syllable, accent a vowel. Expressing facial displays as elaborations of baseline facial structures, we furrow the brow, squint the eyes, pull the corners of the lips down, nod or shake the head, look up, look down with the eyes, and so forth. These are only simple examples, leaving a great deal of detail and complexity to be explored and explained. We believe, however, they demonstrate how CG principles can be applied not just to semantic structure but also to phonological structure.

We have mentioned that grammatical classes are defined in symbolic terms in CG. Conceptual archetypes play a central role here. We have seen that two basic conceptual archetypes are a physical object, and energetic interactions or events. Langacker (Reference Langacker2008, p. 104) describes how these form the basis at the prototype level of noun and verb grammatical categories. For nouns:

1. 1. A physical object is composed of material substance.

2. 2. We think of an object as residing primarily in space, where it is bounded and has its own location.

3. 3. In time, on the other hand, an object may persist indefinitely, and it is not thought of as having any particular location in this domain.

4. 4. An object is conceptually autonomous, in the sense that we can conceptualize it independently of its participation in any event.

For verbs:

1. 1. An energetic interaction is not itself material, consisting instead of change and the transfer of energy.

2. 2. Thus an event resides primarily in time; it is temporally bounded and has its own location in time.

3. 3. By contrast, an event’s location in space is more diffuse and also derivative, as it depends on the locations of its participants.

4. 4. This is so because an event is conceptually dependent; it cannot be conceptualized without conceptualizing the participants who interact to constitute it.

We emphasize that experientially grounded archetypes account for the prototype characterizations of nouns and verbs. Characterizations that are valid for all instances must occur at the schematic level. Certain general cognitive abilities are recruited to arrive at these schemas. For nouns, the schematic characterization is called a thing (used in a specific technical sense), which is the result of grouping and reification (reification allows us to conceptually manipulate a group of entities as a unitary higher-order structure). For verbs, the schematic characterization relies on two cognitive abilities: the capacity for apprehending relationships and the ability to track relationships through time (Langacker Reference Langacker2008, p. 108).

We have seen that many of the properties of language, such as autonomy-dependence, schematicity, and complexity apply to the semantic pole and the phonological pole of symbolic structures. The same is true for conceptual archetypes, which emerge from our embodied physical interactions with the world, including the social world of interacting with others. By saying that they are embodied, we mean that these conceptual archetypes are built up from perceptual and motor interactions with physical objects in the world. We see objects moving, changing their locations, and interacting energetically. Thus, while these experiences give rise to conceptualizations, they are grounded in physical perceptual abilities (hearing, vision, touch, etc.) and in physical motor abilities. This is, in a very general sense, the phonological level – the level at which we produce the physical actions that result in signals perceptible by others.

We emphasize this view because it is this embodied, perceptual, motoric, physical level in which signed language phonology occurs: things acting, hands changing shapes, hands moving and changing locations. In a very real sense, the raw material of sign phonology is one of the experiential sources of conceptual archetypes; or, to put it the other way around, sign phonology is an expressive manifestation of conceptual archetypes.

3.3 Cognitive Models

Cognitive grammar relies on several cognitive models incorporating and building on conceptual archetypes to account for various aspects of language and grammar. We will describe some of these conceptual archetypes and cognitive models in more detail in later sections as we show how they apply to the analysis of signed languages. Here, we briefly describe a few of the cognitive models used in the analysis of grammar.

1. 1. Stage Model: a cognitive model which characterizes a fundamental aspect of our perceptual experience: observing external events which involve the interaction of participants within a setting (Langacker Reference Langacker1991b, p. 284). The stage model evokes the visual experience of watching a play take place on a stage. The maximal scope of our visual attention includes the entire stage; viewers focus their limited visual attention on the actors who occupy locations on stage and interact with each other. The viewers in the audience remain off-stage. As the central action onstage, the participants and their interaction define the immediate scope. Within the area onstage, attention is directed at a particular element, called the profile.

2. 2. Billiard Ball Model: closely related to the stage model and the global setting conceptual archetype is a model in which discrete objects move through space and interact energetically. Their motion may be driven by some external force, or by some internal force; in either case, energy is transferred by forceful physical contact from the moving object to the impacted object (Langacker Reference Langacker1991b). This model is the basis for a conceptual account of clause structure.

3. 3. Canonical Event Model: this cognitive model (Figure 3) combines elements of the stage model and the billiard ball model. The cognitive event model combines the viewing arrangement of the stage model with a viewer observing the onstage action, with the conception of these onstage participants interacting and transmitting energy (Langacker Reference Langacker1991b).

4. 4. Reality Model: a model which the world unfolds in a certain way out of all the conceivable ways. Reality is the history of occurrences. As this is a conceptual model, each person has his or her own conception of reality. The reality model plays a significant role in a cognitive account of modals (Langacker Reference Langacker, Marín-Arrese, Carretero, Hita and Van der Auwera2013).

5. 5. Control Cycle: the control cycle describes a ubiquitous phenomenon consisting of four phases. The control cycle may be manifest in several domains, including physical, mental, perceptual, and linguistic. The best way to understand the control cycle is with a simple physical example: a frog catching a fly. In the first phase, the frog is at rest; this is the baseline phase, and the frog is an actor, having various aspects of its life under control. This is called the frog’s dominion. In the second, some target, in this case the fly, enters the frog’s field (in this example, the frog’s perceptual field), resulting in a state of tension. In the third phase, the frog undertakes some action: it catches and ingests the fly. As a result, the frog is once again in a period of stasis or rest in the fourth phase. The control cycle is manifest in two ways. We have already introduced the first way in the fundamental principle of making sense or epistemic control, which pertains to our knowledge of the world. The second way that the control cycle is manifest is in effective control – our effort to have some effect on or to influence the world. Of course, humans are entities in the world. Effective control is not only the effort to influence others, but ourselves, and so effective control extends to mentally or physically adapting to the world. The control cycle is manifest in several ways in language, including the grammatical expression of possession, modality (both root and epistemic), the expression of assertions, and so forth. We will apply the control cycle to the analysis of facial displays in signed languages in Section 6.

6. 6. Reference Points: reference points are a general cognitive ability that are manifest in all aspects of our experience. Reference points “invoke the conception of one entity for purposes of establishing mental contact with another” (Langacker Reference Langacker1993, p. 5). For example, if we wish to direct someone’s attention to a location we might say, “Do you know where the university library is? Well, the linguistics department is directly across from there.” The library serves as a reference point for mentally locating the desired target of attention, the linguistics department. Reference points appear in possessives, topic constructions, relative clauses, antecedent-anaphor relations, metonymy, and in our analysis of pointing constructions in Section 4.1.

Figure 3 Canonical event model

4.1 Nominal Grounding

Nominal grounding may be understood as a kind of conceptual pointing: the signer or speaker uses various linguistic resources such as demonstratives to conceptually direct the interlocutor’s attention to – point to – the intended referent. Langacker (Reference Langacker2016b) considers pointing to be a kind of linguistic symbol and suggests that an act of pointing is a good point of departure for understanding definite nominal grounding, observing that “a demonstrative constitutes a kind of mental pointing (often accompanied by a physical pointing gesture)” (Langacker Reference Langacker2009, p. 121). Figure 4, from Langacker (Reference Langacker2016b), depicts the CG characterization of a prototypical act of pointing.

Figure 4 Pointing in cognitive grammar

(Langacker, Reference Langacker2016b, used with permission)

In the diagram, G is the actual ground in the current speech event. S and H are the speaker and hearer, where these terms are extended to mean the interlocutors producing (S) and perceiving (H) the pointing. The current discourse environment (or discourse space, CDS) includes the visually accessible immediate physical context. This onstage (OS) region contains a number of things (the circles) which could be singled out by pointing. The solid arrow represents the pointing finger directed at FOC, the focus of attention. The act of pointing has directive force (the double arrow), instructing the addressee to follow, both visually and conceptually, its direction. As a result both interlocutors focus their attention on the same entity, the intended referent. The focused referent is shown in bold to indicate that it is the entity that is profiled.

Pointing Is a Construction

We have seen that conceptual archetypes may range from simple concepts such as a physical object to more complex scenes such as the global setting archetype. Pointing is a complex conceptual archetype composed of component elements. The seventh-century Chinese Buddhist monk Huineng observed that “when the wise man points at the moon, the fool looks at the finger.” We see here the two components of pointing: the moon, which is the focus of attention, and the finger, which directs attention to the moon but is not (except for fools) the focus of attention.

These two components of pointing have been made explicit in a cognitive linguistic account of pointing (Martínez & Wilcox Reference Martínez and Wilcox2019; Wilcox & Occhino Reference Wilcox and Occhino2016). This analysis reveals that pointing is a construction consisting of two component symbolic structures – the device or means of directing attention, which carries with it directive force, and the focus of attention. These two components are termed a pointing device and a Place,Footnote ⁶ each consisting of a form and a meaning. Figure 5 depicts the two component symbolic structures and the (bolded) composite pointing construction. Ellipses in the phonological pole of the pointing device indicate schematicity, subsuming, for example, index finger, hand, eye gaze, body orientation, or even a pointing instrument such as a pencil. The only phonological specification of the pointing device is that it has to be capable of directing attention by pointing. The pointing device and its directive force instruct the addressee to follow its direction, so that both participants in the communicative event focus their conceptual attention on the same entity, the semantic pole of the Place symbolic structure. The only specification of phonological pole of Place is a spatial location (LOC). The schematic semantic pole of Place is “thing” – something conceived as a single entity.

Figure 5 Pointing construction

Places arise through the process of schematization acting on our perceptual and experiential world. Consider the experience of two friends having cappuccinos in their favorite coffee shop. One person points to the coffee cup and says, “Look at this cup. Isn’t it cute?” In pointing, he directs attention to an object in a location, the experiential archetype for a Place. We can extend this account to encompass many different situations. We point to objects that are present (and known to be present) in the physical environment but not visible, such as a familiar building that is too far away or blocked by other buildings to be seen, or a distant mountain the view of which is blocked by trees. We may point to a location that was formerly occupied by some object that has not been present in that location for years, decades, or even centuries. For example, upon visiting a childhood home with a friend, one might say, “This [pointing to a location in the living room] is where my piano was when I was a child.” In signed languages, as we will see, we can point to entities that exist but are not and never were in the current spatial environment, a type of anaphoric reference. We can also point to virtual or abstract entities, such as ideas, schools of thought, emotions, and so forth. This latter use of Places is especially productive in contrastive constructions (Winston Reference Winston, Emmorey and Reilly1995).

As a signer perceives and produces more of these usage events, she abstracts away from the semantic and spatial specifics of any particular entity and its location, extracting what is common across multiple experiences to arrive at a schematic Place symbolic structure. Schematization operates at both poles of the symbolic structure – semantic and phonological. The spatial location of an entity is, for signed languages, an aspect of its phonology. The location of the signer’s nose, eyes, and ears is irrefutably specified phonologically. By extension, the spatial location of the cappuccino cup, a mountain, and even the remembered piano is for a signer a phonological location.

Thus, Places are abstracted from actual usage events – in this case, the archetypal usage event of pointing to a physical object in a spatial location. Being schematic, a Place has neither a specific meaning nor a specific spatial location; rather, Places are symbolic structures that associate a schematic meaning with a schematic phonological (spatial) location. In a usage event, the schematic meaning and the schematic location are instantiated, resulting in a fully contextualized Place. In a usage event, a signer directs attention to the phonological pole of the Place, its location, and because of the association between the phonological pole and the semantic pole of symbolic structures, the signer thereby directs attention to the intended referent.

An important aspect of schematic meanings is that they are immanent in their instantiations. Immanent meanings ‘lie within’ those of instantiating expressions, which elaborate them (‘flesh them out’) in their own individual ways” (Langacker Reference Langacker2009, p. 14). The more schematic concept of liquid is immanent within the specific instance water and cappuccino. Immanence is a central notion in CG present in a host of expressions, including possessives, reference point phenomena, epistemic versus root modality, and grammatical categories. Immanence also pertains to cognitive abilities. Conceiving an object moving through space requires mental scanning through time in the spatial domain. Langacker (Reference Langacker1993, p. 4) points out that we are usually not aware of the mental scanning as a separate mental experience “precisely because it is immanent in (or manifested through) the richer, more contentful conception of spatial motion by a physical object.” Just as mental scanning is immanent in the conception of an object moving through space, schematic Places are immanent within the objects with which they spatially coincide.

Because of the range of entities that we point to – physically and visually present cups, missing cups, unseen but still physically existing buildings, the memory of the location of a childhood piano, and more – we can posit a continuum of Places (Figure 6).

Figure 6 Places continuum

Points along the continuum can be described with respect to various dimensions. One dimension is spatial, temporal, and conceptual immediacy. Spatially present objects fall at the left end of the continuum; they are real, and visual perception provides direct access to the physical entity and its Place. Temporal immediacy pertains to the present moment. Temporally distant, nonimmediate, experiences require progressively greater conceptual resources. Experiences such as the missing cappuccino cup require recall: one must know and remember that the cup was present in the current discourse environment. The cognitive load here is fairly minimal. The same is true to an extent for the building or mountain that cannot be seen: visual perception is no longer available, requiring memory and a mental projection that the entity is still in existence. Naturally time is a factor as well: the longer the time between present and when the cup, the mountain, or the building was seen requires great memory. The childhood piano falls farther along the continuum. Pointing to the location occupied by the childhood piano places an even greater load on memory and mental projection. At the far right are virtual and imaginary entities (e.g., theories) and real but nonpresent entities which exist independently of any particular spatial location, but are used in anaphoric pointing constructions. These and other factors determine in a complex interaction the cognitive resources required.

From the viewpoint of how humans perceive and construct concepts of our experiential world, locations do not exist. We see objects that occupy locations. This is the experiential basis of the conceptual archetype object in a location. This is also the perceptual and experiential basis of Places. The baseline Place, perceptually and developmentally, is a perceptible, physical object. With elaboration requiring increasing cognitive resources, a Place can be a remembered entity, a virtual (real but not present) entity, or an imagined entity that we mentally place at some location – by pointing to it or by placing it at the location; the latter operation is also a common grammatical operation used in signed languages, as we will see in Section 4.2.

One unification suggested by this account is that the set of Places along the continuum in Figure 6 subsumes the traditional distinction made between deixis and anaphora. Talmy (Reference Talmy2018, p. 1) writes that “Broadly, an anaphoric referent is an element of the current discourse, whereas a deictic referent is outside the discourse in the spatiotemporal surroundings. This is a distinction made between the lexical and the physical, one that has traditionally led to distinct theoretical treatments of the corresponding referents.” Talmy proposes an alternative: “language engages the same cognitive system to single out a referent whether it is speech-internal or speech-external.” Like Talmy, we suggest that these two domains of reference are manifestations of the same conceptual system. Further, because signed languages are expressed in space, and in a great many constructions both deictic referents and anaphoric referents are associated with locations in space, we find no need to impose a distinction between these types of locations. Conceptually and phonologically they are part of the same linguistic system and manifest as Place symbolic structures. We will take up this matter again in Section 7.

To summarize, Places are symbolic structures consisting of a schematic semantic pole (the most schematic meaning of Place is “thing”) and a schematic phonological pole (a location in space). The baseline Place is a perceptible physical object. Through our experience with the world we construct networks of Places with varying degrees of semantic (what is the thing?) and phonological (where is it located?) schematicity/specificity. Cognitive abilities such as memory and imagination operate to yield elaborations of this baseline situation, with increased detachment from perception of the experiential world requiring increased conceptual resources. By using these general cognitive abilities and the linguistic resource of Places, signers are able to semantically instantiate and phonologically direct attention to an unlimited number of Places that refer to distinct discourse entities. Obviously, this analysis is valid for gesturers as well. This does not mean that signs are gestures, just as a dynamic systems model can account for both linguistic and nonlinguistic skilled movements (Hawkins Reference Hawkins, Docherty and Ladd1992; Saltzman & Kelso Reference Saltzman and Scott Kelso2009). We will discuss the issue of what counts as a sign and what counts as a gesture in Section 7.

Pointing Device

The other symbolic component of pointing constructions is the pointing device. In previous work we have characterized the schematic meaning of pointing devices as directing attention (Wilcox & Occhino, Reference Wilcox and Occhino2016). How is this accomplished? Conceptually, a pointing construction selects a particular referent from a pool of candidate entities in our mental universe. It does so by mentally pointing to the selected entity (the arrow in Figure 7), thereby profiling (bolded) it as the focus of attention and indicating its status relative to the ground (G). This candidate pool (large circle) and the potential referents (small circles) are elements of the semantic pole of the pointing device. Semantically, the pointing device does not make reference to a specific referent; rather, the selected referent is a schematic dependent structure internal to the pointing device’s semantic pole (cross-hatching in Figure 7 indicates schematicity). This schematic structure must be elaborated, and this is the function of the semantic pole of a Place, depicted by the right portion of Figure 7.

Figure 7 Pointing device semantic pole

The pointing device is a grounding element that indicates the epistemic status of the referent Place in relation to the ground. The grounding function of the pointing device is further elaborated in Figure 8. The ground (G) is offstage, although part of the conceptual maximal scope. The pointing device as grounding element directs attention to a schematic dependent semantic structure and puts it onstage as the focus of attention in the immediate scope. The Place component corresponds to (dotted line) and elaborates (arrow) the schematic focus of attention (schematicity here indicated by gray hashing). The composite level is shown at the top of Figure 8. The pointing device as grounding element directs attention to and profiles the specific referent instantiated by the Place.

Figure 8 Pointing device as grounding element

We have simplified the description in a way that requires a more detailed explanation. When we say that the semantic pole of Place instantiates and elaborates the schematic element profiled in the semantic pole of the pointing device, the way this happens depends on the context of use. When we point to a physically present cappuccino cup, the semantic pole of Place is directly instantiated by the cup. In other uses, the semantic pole of Place is itself also schematic (“thing”), and has to be elaborated elsewhere in the discourse. Thus, the bottom portion of Figure 8 may extend to other expressions in the discourse, and the final specification of the referent will depend on a chain of correspondences. This is further exemplified by proxy-antecedent constructions (see Figure 11).

In summary, the semantic pole of the pointing device makes internal reference to a dependent element, mentally directing attention to it. This dependent element is elaborated by the semantic pole of a Place symbolic structure. The referent is either directly specified, or it is specified elsewhere in the discourse through a chain of correspondences. The pointing device is a grounding element, specifying the profiled referent schematically and putting it onstage as the profiled focus of attention. The referent, the semantic pole of the Place, is what the pointing device profiles in the composite pointing construction.

Pointing Constructions in Signed Languages

Pointing is a ubiquitous feature of the world’s signed languages. Pointing signs function as “deictic and anaphoric pronouns, possessive and reflexive pronouns, demonstratives, locatives, determiners, body part labels, and verb agreement” (Meier & Lillo-Martin Reference Meier and Lillo-Martin2013, p. 151). Here we will examine anaphoric pronouns.

One aspect of grounding that pronouns must accomplish is locating the referent in relation to the ground. Another factor is singling out a unique referent. In the case of first- and second-person pronouns, both of these factors are intrinsic, since they single out referents inherent to the ground, the speaker/signer and the addressee. A more difficult problem is raised by third-person pronouns. It is possible that many referents are available in the current discourse. To be used felicitously, a third-person anaphoric pronoun must uniquely identify an antecedent referent. For example, if someone says My new Honda CR-V is a hybrid. It is the first hybrid I have ever owned, the interlocutor must identify the full nominal in the prior discourse to which it refers.

The CG description of the semantics of pronominal anaphora relies on conceptual reference point relationships (Langacker, Reference Langacker1993; Langacker Reference Langacker2000; Van Hoek Reference Van Hoek1997). The reference point relationship is shown in Figure 9, in which C is the conceptualizer; R is the reference point, a salient entity in the current discourse space; T is the target structure to which R provides access; and D is the dominion, the set of potential targets. A spatial example of a reference point occurs when someone says, “The linguistics department is directly across from the library”; the reference point is the library and the target is the linguistics department. The dominion is all the possible targets that may be located relative to the reference point.

Figure 9 Reference point

(Wilcox & Occhino, Reference Wilcox and Occhino2016)

For anaphoric pronouns (Figure 10) the semantic pole of a pronoun profiles a schematic thing (indicated by ellipses in the target). It also incorporates the assumption that the speech act participants have mental access to the intended referent, the full nominal antecedent which serves as the pronoun’s reference point. Mental access is provided by the reference point relationship (indicated by the dashed arrow): the pronoun target is in the dominion of the reference point antecedent, which is presumed to be salient and accessible to the interlocutors in the current discourse context. From the addressee’s perspective, an anaphor gives instructions to seek a prominent entity which serves as the target’s reference in the current discourse context.

Figure 10 Antecedent-anaphor relation

(from Wilcox & Occhino, Reference Wilcox and Occhino2016)

The reference point analysis of the relationship between anaphoric pronouns and antecedents applies to signed languages as well, with one exception. In the Honda CR-V example, there are two elements: the antecedent (my Honda CR-V) and the anaphor (it). A similar antecedent-anaphor construction in a signed language often requires three elements: the antecedent, a proxy antecedent, and the anaphoric pronoun.

For example, in Australian Sign Language (Auslan), a narrator signs a noun, TEACHER, which is then followed by a pointing sign directed to a particular location on the right of the signer, establishing an association between the referent and the right side of the signing space. Later in the narrative, the signer again points to the location on the right to refer to the teacher. Expressions such as this are common in signed languages and indicate antecedent anaphor relations (Johnston & Schembri Reference Johnston and Schembri2007, pp. 271–272).

In our CG analysis, a Place symbolic structure conventionally associates a referent with a phonological location in signing space. In the Auslan example, this is a location on the right side of the signer. The semantic pole of the Place symbolic structure profiles a schematic thing, which is associated with and elaborated by the semantic pole of the antecedent TEACHER. This Place thus serves as a proxy antecedent for TEACHER. In proxy-antecedent constructions the signer does not point directly to the lexical antecedent TEACHER, but to the phonological location of the proxy antecedent Place. In Figure 11, the semantic poles of the antecedent (TEACHER) and the proxy antecedent are connected by correspondence lines because the two are the same entity: the antecedent TEACHER instantiates the schematic semantic pole of the proxy antecedent Place. Later in the narrative, the signer again points to this proxy-antecedent Place; here the pointing construction designates a pronominal anaphor. The entity designated by the semantic pole of the Place component of this pointing construction, and the entity that semantically instantiated the Place component of the prior proxy antecedent – the lexical sign TEACHER – are conceptually mapped to the same entity as indicated by the correspondence lines. The anaphor, the proxy antecedent, and the antecedent (TEACHER) conceptually overlap: they are co-referential.

Figure 11 Proxy-antecedent anaphor construction

(Martínez & Wilcox, Reference Martínez and Wilcox2019)

Proxy-antecedent constructions exhibit a unique feature only partially seen in spoken language antecedent-anaphor constructions. In most reference point constructions, the target is distinct from the reference point. In the previous example of the linguistics department and the library, the reference point (library) and target (linguistics department) are distinct entities. In anaphor-antecedent constructions, however, the reference point and the target are co-referential – my Honda CR-V and it, and TEACHER and the second pointing construction in Figure 11 are conceptually the same entity. Langacker (Reference Langacker2000, p. 238) describes this as a degenerate reference point relationship: the reference point (antecedent) and the target (pronoun) collapse to a single entity in conceptual space (indicated by the correspondence line connecting the two in Figure 11). In signed languages, this conceptual degeneracy is expressed symbolically by phonological degeneracy: the phonological location of the proxy antecedent and the phonological location of anaphor – both of which are Place structures – coincide.

The same phonological location across two or more symbolic structures is a common grammatical device used in many signed languages. This is a unique manifestation of Behaghel’s law, the principle that elements that belong together semantically tend to be placed together syntactically. Langacker (Reference Langacker1987) observes that the most straightforward kind of togetherness is linear juxtaposition, where one structure is placed immediately adjacent to another structure.Footnote ⁷ That is true for spoken languages, but for signed languages an even more straightforward kind of togetherness is possible: placement of the phonological poles of two structures at the same spatial location.

Placing

In addition to being components in pointing constructions, Places are also components in placing expressions. The term placing was introduced by Clark (Reference Clark and Kita2003, p. 185), who identified pointing and placing as two forms of gestural indicating, that is, of creating indexes for things. In pointing, speakers direct their addressee’s attention to the object they are indicating. In placing, “speakers try to place the object they are indicating so that it falls within the addressees’ focus of attention” (Reference Clark and Kita2003, p. 187). Clark offered the example of purchasing items in a drugstore: a customer identifies the items to be purchased by placing them on the counter, and then identifies the payment by placing cash on the counter.

Martínez and Wilcox (Reference Martínez and Wilcox2019) extended the concept of placing in the context of signed languages. First, they observed that in signed language constructions, signs are communicative objects that can be placed in spatial locations. Second, they identified two types of placing: create-placing, in which a new Place is created, and recruit-placing, in which the signer recruits an existing Place. Figure 12 depicts a generic placing construction. S is the signer, I is the interlocutor, and G is the ground. The bold line with ball end indicates the act of placing. The dashed line with a magnet end indicates the subtle distinction between pointing and create-placing: rather than directing attention, create-placing locates an entity so that it falls within the addressee’s focus of attention, and thus attracts the attention of the interlocutor to the Place.

Figure 12 Placing

(Martínez & Wilcox, Reference Martínez and Wilcox2019)

An example of placing in Argentine Sign Language (LSA) is given in Martínez and Wilcox (Reference Martínez and Wilcox2019). The signer introduces the biography of José de San Martín, a hero of the independence of Argentina, Chile, and Peru. At the beginning of the discourse a noun, PERSON, is create-placed at the right side of the signer, creating a new Place (Figure 13). The schematic semantic pole of the Place is elaborated by the semantic pole of PERSON, and the schematic phonological pole of the Place is elaborated by the location on the right of the signer. Constructions specific to a signed language often specify phonological locations for new Places, for example, perhaps specifying that new discourse entities are introduced on the signer’s dominant side as we see here. The placing construction locates PERSON as a newly introduced referent intersubjectively identified and accessible to the interlocutors, and thus a grounded (but still indefinite) nominal.

Figure 13 Create-placed PERSON and Point

(Martínez & Wilcox, Reference Martínez and Wilcox2019)

Once the referent Place is created, the signer is able to refer to it in subsequent discourse. For instance, the sign RENOWNED (Figure 13) incorporates the nominal referent “person” as a participant of the adjectival relation by directing the sign toward the “person” Place. RENOWNED profiles one focal participant and associates the property of being famous with San Martín. Later in the discourse, the signer refers anaphorically to the same referent with pointing constructions (glossed as POINT(location)) directed to the Place instantiated by San Martín.

Pointing and placing are used to ground and track nominals at multiple levels of structure, from clauses to larger constructions. An example of more complex use of pointing comes from a video in LSA of the official account of the Argentine Deaf Movement or, in Spanish, Movimiento Argentino de Sordos (MAS) in support of a bill recognizing Argentine Sign Language. Among other political actions, the National Association of the Deaf in Argentina (CAS, Confederación Argentina de Sordos) organized an event at the National Congress building that gathered together more than 5,000 people. In the video, released two days before the event, two deaf leaders of the Movement, Alejandro Makotrinsky and Pablo Lemmo, make the point that the MAS movement should not attempt to label or classify hearing people as inherently wrong or bad.

Pablo then introduces what will become the overall topic: widespread ideology leads hearing people to believe that deaf people are mentally challenged, not equal, or are deaf-mute. He does this by first signing IDEOLOGY, a two-handed sign with a location at the head (Figure 14). Then, while his nondominant hand is still in the head location for IDEOLOGY, he begins to point to it with his dominant hand. More precisely, by the time Pablo completes the pointing action his nondominant hand has moved down to a neutral position (as seen in the second panel of Figure 14), and so the sign IDEOLOGY is no longer present (like the missing cappuccino cup), and so he is pointing to the spatial location formerly occupied by IDEOLOGY. This pointing construction grounds IDEOLOGY as a nominal and creates an ideology-Place. However, rather than simply pointing up to the sign (since its location is high in signing space), Pablo reaches his pointing arm up above the sign and points down at it. Downward pointing in LSA, and possibly in other signed languages, has been shown to express proximal meaning (Martínez & Wilcox, Reference Martínez and Wilcox2019). Here, Pablo wants to create a sense of conceptual closeness with the concept of ideology, emphasizing that “this thing we call ideology, this is what we have to attend to and focus on.” Pablo articulates the pointing device with several forceful reduplicated movements. It has also been shown that increased directive force can be expressed in various ways in LSA (Martínez & Wilcox, Reference Martínez and Wilcox2019), including reduplicated movements of the pointing device directed at the referent Place. Pablo then continues.

Figure 14 IDEOLOGY pointing and recruit-placing

(Martínez & Wilcox, Reference Martínez and Wilcox2019)

To express the idea that the goal is to change society’s ideology pertaining to deaf people, Pablo recruit-places the sign CHANGE in the newly created ideology-Place (Figure 14). In this construction, CHANGE is unspecified for phonological location; the schematic location of CHANGE is elaborated by the location of the ideology-Place, which has been previously elaborated by the phonological location of IDEOLOGY. CHANGE profiles an action chain, which includes an unexpressed agent and a theme, the changed entity. In this construction, the theme is elaborated by the ideology-Place, which in turn conceptually maps to IDEOLOGY.

Figure 15 shows IDEOLOGY (a) and the ideology-Place (b) created by the pointing construction. IDEOLOGY is a lexical noun with full phonological specification (HC is hand configuration, MOV is movement), including the head location. IDEOLOGY is grounded as a full nominal by the proximal (downward directed) pointing device, which also creates an ideology-Place. IDEOLOGY thus corresponds to and elaborates the schematic Place. CHANGE is then recruit-placed at the location of the ideology-Place, indicating that “ideology” is the theme (TH), the changed entity. CHANGE is shown as a construction consisting of two symbolic structures: the action chain process “change” (double arrow) and the theme “ideology.” CHANGE incorporates a schematic symbolic substructure (indicated by the rectangle enclosing the theme (TH) and the location (LOC). The semantic pole of this symbolic substructure specifies the theme. The schematic phonological pole is elaborated by the placing construction. Once again, we see double overlap of conceptual mapping and phonological mapping of Place symbolic structures.

Figure 15 CHANGE recruit-placed

(Martínez & Wilcox, Reference Martínez and Wilcox2019)

4.2 Place and Placing in Agreement Constructions

Kibrik (Reference Kibrik2019, p. 76) has observed that “In modern linguistics, ‘agreement’ has developed into an entangled and incoherent notion.” The entanglement and incoherence extends to and is amplified in discussions of agreement in signed languages (Cysouw Reference Cysouw2011; Fischer & Gough, Reference Fischer and Gough1978; Liddell Reference Liddell, Emmorey and Lane2000b; Lillo-Martin & Meier Reference Lillo-Martin and Richard2011; Quer Reference Quer2011; Schembri et al. Reference Schembri, Cormier and Fenlon2018). Kibrik categorizes approaches to agreement in spoken languages into two types: a form-to-form approach, and a cognition-to-form-mapping approach. The traditional form-to-form approach claims that an agreement feature originates in one linguistic element, the controller, and is copied onto another one, the target. As Kibrik points out, the form-to-form approach presents several difficulties. For example, sometimes the agreement marker has nothing to agree with, or expresses entirely different information (Barlow Reference Barlow1999; Croft Reference Croft, Bakker and Haspelmath2013; Langacker Reference Langacker2008).

In the cognition-to-form-mapping approach, agreement features are associated with referents in the cognitive representation. They serve to specify referents and are thus referential features which are mapped onto various sites in linguistic structure. The roots of agreement, according to Kibrik (Reference Kibrik2019, p. 37), “lie in cognitively motivated discourse processes associated with reference.” Such an approach is also offered by Croft, who suggests it is possible to treat agreement affixes as expressing a symbolic relation (rather than a syntactic relation) indexing the referent, and thus to treat agreement as “double indexation” (Croft Reference Croft2001, p. 229).

Agreement in CG is analyzed as multiple symbolization: “That is, information about some entity is symbolized by more than one component structure within the same symbolic assembly and thus has multiple manifestations in a single complex expression” (Langacker Reference Langacker2008, p. 188). Agreement in this view is “simply a matter of the same information being symbolized in multiple places. As such, it is just a special case of conceptual overlap, which is characteristic of all grammatical constructions” (Langacker Reference Langacker2008, p. 347). We rely on the CG account and analyze agreement as a special type of conceptual overlap: the same information is symbolized in multiple manifestations of symbolic structures that map to the same cognitive referent. The predominant way conceptual overlap is symbolized in these constructions is with phonological overlap: the double overlap we have already described.

We have already seen examples of multiple symbolization in the use of Place. In Figure 13, the sign RENOWNED incorporates the referent “person” as a participant of the adjectival relation by directing the sign toward the “person” Place, profiling the focal participant, San Martín, who had been associated with that Place in a previous discourse frame. The sign RENOWNED, in our analysis, contains a schematic symbolic substructure. Semantically, the substructure is specified as the antecedent of “renowned.” The phonological pole of the substructure is specified as the end location of the movement path. In the example, this schematic substructure is phonologically elaborated by being placed at the location of the San Martín-Place. “Agreement” is thus manifest as two symbolic structures, or multiple symbolization: the schematic symbolic structure internal to RENOWN maps to the San Martín-Place symbolic structure, which elaborates its semantic and phonological poles.

A verbal example appears in Figure 14. Pablo has created an ideology-Place with a pointing construction. He then placed the sign CHANGE at the ideology-Place. In this construction, the semantic pole of CHANGE contains a schematic substructure specifying the theme of the verb “change” and a schematic phonological pole, both of which must be instantiated. In the placing construction, the schematic semantic substructure is instantiated by the semantic pole of the ideology-Place, which in turn corresponds to “ideology.” The schematic phonological location is instantiated by the phonological location of the ideology-Place, which corresponds to the phonological location of “ideology.” Again, we find that multiple symbolic Place structures map to the same conceptual entity, “ideology.”

In the same Movimiento Argentino de Sordos video, Pablo produces another agreement form. Continuing his narrative, Pablo explains that because of their ideology, hearing people regard deaf people as mentally challenged, not equal to hearing people, mute, and incapable. Prior to this segment, Pablo had placed the sign CHANGE directed toward himself, recruiting his body as a Place symbolic structure to represent people in general; in this context, it is clear that he refers to hearing people who have negative attitudes toward deaf people. Even though he is actually deaf, Pablo uses this argumentative strategy because he wants to convince hearing people to change their negative view of deaf people without directly confronting them.

Pablo again produces the sign PERSON with his nondominant hand, placing the sign at the spatial location occupied by Alejandro – in other words, at the Alejandro-Place. This placing construction is a component in a larger, simultaneous construction (Figure 16). While Pablo continues to hold the placed sign PERSON, he signs DEAF TO-SEE DEAF with his dominant hand. The verb TO-SEE is produced with a path movement from Pablo toward Alejandro. Pablo then lists the negative characteristics that hearing people perceive deaf people as having. TO-SEE in the sense used here is not a perceptual verb; rather, it is a verb of cognitive activity. TO-SEE means “to regard as” or “to think of,” and in this use it expresses the cognitive activity of hearing people categorizing deaf people.

Figure 16 Simultaneous placing construction

The lexical sign PERSON is a type specification. When placed, it integrates the “person” type with the semantic pole of the Alejandro-Place. However, the semantic pole of this Place is not Alejandro. Rather, Alejandro’s Place serves as a reference point, affording mental access to a dominion of targets. We have previously shown that reference point constructions in signed languages are used to associate characteristics or properties as reference point targets with referents (Martínez & Wilcox Reference Martínez and Wilcox2019), in this case, characteristics associated with Alejandro. The targets could be Alejandro’s gender, his hair color, his clothing, or any number of other characteristics. In this situation, the most salient target is Alejandro’s deafness. This characteristic is explicitly mentioned when Pablo signs DEAF TO-SEE DEAF. As shown in the lower portion in Figure 17, the PERSON component integrates with the Alejandro-Place component, specifically the target “deaf” of the reference point, to create the composite construction “deaf people.”

Figure 17 Simultaneous placing construction

This composite construction is then a component in a higher-level construction that integrates “deaf people” with TO-SEE. TO-SEE has two schematic symbolic substructures: the categorizer and the object of categorization. These two substructures are elaborated by Place structures. The semantic element of the categorizer substructure is elaborated in the prior discourse frame when Pablo uses the placing construction to present himself as a hearing person. Pablo is making the point that hearing people are doing the categorization. The semantic element of the second substructure, the object of categorization, is elaborated by the composite construction “deaf people,” producing the complex construction “hearing people see deaf people” as incapable, and so on. The middle portion of Figure 17 shows that the component “deaf people” conceptually elaborates and corresponds to the schematic object of conceptualization element (O) of TO-SEE. The component “hearing people” conceptually corresponds to and elaborates the schematic conceptualizer (C), and integrates with the component “deaf people” to yield the final composite construction “hearing people ‘see’ deaf people.”

The phonological side of these constructions shows comparable complexity. The relevant issue is the location of the Place structures. The phonological location of PERSON is schematic, allowing it to be placed. When PERSON is placed in this context, the phonological location of the Alejandro-Place elaborates the schematic phonological location of PERSON. The phonological poles of the two schematic substructures of TO-SEE are also elaborated by Places. The Pablo-Place (representing hearing people) elaborates the phonological location of the first schematic substructure, the categorizer. The schematic phonological location of the second substructure, the object of categorization, is elaborated by the phonological location of the Alejandro-Place (representing deaf people).

Observe that both groups are virtual: no actual hearing or deaf people are involved, although both are expressed by Places occupied by real people in the spatial environment. The virtual deaf people are evoked by Alejandro’s Place when integrated with the type specification PERSON. The conceptualizer of TO-SEE is also a Place (Pablo’s) created by the placing construction in the previous discourse frame to create virtual hearing people. Thus, Places associated with physical entities – Pablo and Alejandro and their spatial locations in the ground – play essential roles in the component structures that form this complex grammatical construction referring to virtual groups of people. This use of Places that correspond to physical entities in the spatial environment has been described for other grammatical expressions in LSA (Martínez & Wilcox Reference Martínez and Wilcox2019; Wilcox & Martínez Reference Wilcox and Martínez2020) and appears to be common to other signed languages as well. That signers can use Places associated with physical entities in the spatial environment is a manifestation of the Place continuum. It is not, however, generally accepted that the grammars of signed languages are capable of incorporating real-world locations in the discourse ground in this way. We will discuss this issue more in Section 7.

A final example of multiple symbolization and conceptual overlap in LSA occurs in a narrative about a famous event in Argentina. The signer says, “This man, Lagomarsino, the one who gave the gun to Nisman … ” The signer first points to a location in front and slightly to her right, creating a Place, which for the moment remains schematic: we don’t know what or who this Place refers to. She then signs MAN, a body-anchored sign produced at the mouth. Next, she places the sign PERSON at the newly created Place. She then fingerspells the name Lagomarsino followed by a relative marker meaning “the one who” directed at the Place. Finally, she signs an agreement verb GIVE, moving from the Place to a location on her left.

Figure 18 diagrams this excerpt. The solid line (a) is the first pointing construction which creates a Place (b). Line (c) represents the placing of PERSON. Line (d) depicts the relative marker directed at the Place. Finally, circle (e) shows the initial location of the verb GIVE (indicated by an arrow), and (f) shows the final location.

Figure 18 Lagomarsino discourse segment

Figure 19 depicts the semantic pole of the structures in this discourse segment. The arrows indicate the appearance of Place structures across the segment. A semantically schematic Place is created by pointing. Body-anchored MAN elaborates the entity type; we now know that the Place refers to a man. PERSON, placed at the man-Place, provides phonological substance to the location of the Place. Fingerspelled “Lagomarsino” fully elaborates the Place: we now know the referent of the Place. The relative marker, recruit-placed at the Lagomarsino-Place, tells us that the referent at this Place is the one who will be doing something. Finally, we learn that Lagomarsino gave a gun (to the person who will be described in the subsequent discourse frame). The initial phonological location of GIVE is placed at the Lagomarsino-Place.

Figure 19 Lagomarsino semantic pole

Figure 20 depicts the semantic and phonological correspondences. The semantic poles of the initial Place, MAN, PERSON, Lagomarsino, the head of the relative markers, and the agent of GIVE correspond – they map to the same conceptual entity. At the bottom of the diagram, the phonological pole correspondences are shown: the initial Place, PERSON, the head of the relative marker, and the agent of GIVE all are produced in the same spatial location. MAN is body-anchored and thus has a different phonological location (near the mouth), and Lagomarsino is fingerspelled at a location in neutral signing space.Footnote ⁸

Figure 20 Lagomarsino semantic and phonological poles

In summary, these examples demonstrate grammatical constructions characterized by multiple symbolization and conceptual overlap. We also see that in these constructions, conceptual overlap is achieved by phonological overlap: the phonological poles across multiple symbolic structures are articulated at the same location in space, the phonological location of a corresponding Place structure. This double overlap for the Lagomarsino example is shown in Figure 21. The initial Place structure (created by a pointing construction), placed PERSON, the relative marker, and the initial symbolic substructure (agent) of GIVE all map to the same entity in conceptual space – Lagomarsino. They also map to the same location in phonological space.

Figure 21 Double overlap in multiple symbolization

Not all uses of GIVE and similar verbs require full elaboration of each schematic substructure. GIVE may appear in constructions where the agent substructure is left unspecified. For example, if a signer reports that he took his car to a shop and received a bill for the repairs, he might sign GIVE so that the initial location is distal from the signer, moving toward and ending at the signer. This would express a passive meaning, “a bill was given to me,” where the agent of “give” is left unspecified and nondeployable in subsequent discourse (Janzen et al. Reference Janzen, O’Dea and Shaffer2001). The initial location is phonologically elaborated by a conventional Place in the grammar of ASL, the distal location. The semantic pole of this distal Place is a schematic “thing” with no further elaboration; when it integrates with the initial substructure of GIVE, which has only semantic specification of agent, the composite symbolic assembly leaves the agent unspecified and defocused – characteristics of passive constructions. The ending substructure is semantically and phonologically elaborated by the signer Place: the signer is the recipient of the bill. This is also an example of multiple symbolization and conceptual overlap. The essential difference between this use and the Lagomarsino use is that in this construction, the initial substructure of GIVE remains semantically unspecified except for its status as agent.

In the next section we show that signs are not the only communicative objects that can be placed. The signer is also a symbolic structure that can be placed, again resulting in conceptual and phonological overlap.

4.3 Placing the Signer in Reported Dialogue

Dialogue in narrative can be presented either as a third-person report (indirect quotation) or as first-person (“direct quotation”) (Chafe Reference Chafe and Tannen1982). Tannen (Reference Tannen and Coulmas1986) points out that what is referred to as reported speech or direct quotation is actually constructed dialogue. Speakers mark these constructed dialogues with certain conventional grammatical constructions, or by taking on the voices of characters by changes in pitch, voice quality, and prosody (Schiffrin Reference Schiffrin1981; Tannen Reference Tannen and Coulmas1986). In doing so, the speaker takes the point of view of a character in the narrative.

Just as speakers have ways of presenting a point of view by taking on the vocal and behavioral qualities of characters, signers use their whole bodies and the space surrounding them to convey viewpoint in reported dialogue. Padden (Reference Padden1986) offers an example, depicted in Figure 22. The signer says, “The husband goes, ‘Really, I didn’t mean it’.” In the first frame the signer faces her actual interlocutor and signs HUSBAND, identifying who will be speaking in the next sequence. The next four frames present the constructed dialogue REALLY ME NOT MEAN “Really, I didn’t mean it” as signed by the husband. To mark the constructed dialogue, the signer shifts her body to the right and directs her eye gaze at the husband’s virtual addressee.

Figure 22 Role shift

(Padden Reference Padden1986)

Constructions such as this that are used for expressing quotations from a character viewpoint have received different names in the sign linguistic literature, such as role shift (Padden Reference Padden1986; Quer Reference Quer, Pfau, Steinbach and Herrmann2016), shifted reference (Engberg-Pedersen Reference Engberg-Pedersen1993), and constructed action (Cormier et al. Reference Cormier, Smith and Zwets2013; Ferrara & Johnston Reference Ferrara and Johnston2014; Metzger Reference Metzger and Lucas1995). Descriptions of these constructions in signed languages tend to include one or more of these phonological features: (i) a change in body orientation (the signer changes orientation, for instance, from front position to a sideways position); (ii) a change in eye gaze direction (the signer tends to break eye contact from the actual addressee to look in another direction, such as the location of virtual character or referent within discourse); (iii) and a change in deixis (the deictic center, the body of the signer, is rearranged to take somebody else’s point of view). As Engberg-Pedersen (Reference Engberg-Pedersen1993) has observed, these constructions also co-occur with some or all of the following semantic features: (i) shifted reference: the use of pronouns to refer to somebody other than the sender/narrator; (ii) shifted attribution of expressive elements: the use of the signer’s face and/or body posture to express emotions or attitudes of somebody other than the sender/narrator in the context of utterance; and (iii) shifted locus: the use of the sender/narrator locus for somebody other than the sender/narrator.

In Section 3 we introduced the stage model. One aspect of the stage model is viewing arrangement, defined as the overall relationship between the “viewers” and the situation “viewed” (Langacker Reference Langacker2008). Within CG, the term “viewers” refers to conceptualizers who apprehend the meanings of linguistic expressions; they are the signer/speaker and the addressee. In a communicative event, the viewers are the speaker/signer and interlocutor. In the default viewing arrangement (Figure 23), the viewer is offstage while the object of perception (P) is onstage (OS) within the overall field of perception. The default viewing arrangement also reflects the asymmetry between what is viewed, the object of conceptualization, and the viewer, the subject of conceptualization (Langacker Reference Langacker1991a). When this asymmetry is fully polarized, as in the default viewing arrangement, the subject of conception is construed with maximal subjectivity, as the viewer fully attending to the object of perception in the objective scene; the viewer is not a part of that scene and has no awareness of self as viewer. In an alternative construal, the viewer moves onstage and may even become the focused object of conceptualization. Expressions which explicitly reference the ground, such as the signer/speaker or the interlocutor, such as I, me, and you, manifest this viewing arrangement. Other aspects of the ground which are normally off-stage can be moved onstage with words such as now or here.

Figure 23 Default viewing arrangement

(Langacker Reference Langacker1991a)

It is important to note that while viewing arrangement is experientially grounded in the embodied link between perception and conception, the stage model and viewing arrangement pertain to conceptualizations: they characterize the meaning of linguistic expressions, the semantic pole of symbolic structures. For both spoken and signed languages, these cognitive models pertain to semantic structure.

In signed language discourse the stage model and viewing arrangement retain their physical, experiential, and perceptual grounding. This is seen when signers present narrative, either in person or more and more frequently on social media: a signer as narrator literally occupies a location on a stage that is visually perceived by the interlocutor as audience. In the default narrative arrangement, the signer and any interactional partners virtually represented through reported speech are on-stage participants, while the audience remains off-stage as viewer. In presenting narratives, the signer as narrator occupies a location; the narrator then assumes character roles in the narrative, interacting with other (virtual) participants. Thus, for signed languages, the stage model and viewing arrangement are aspects of both meaning and form. The signer, the signer’s location “on stage,” and the relation of the signer to the viewers have phonological significance.

We have shown that the signer may function as a symbolic structure. When Pablo places the sign PERSON directed at Alejandro, he recruits Alejandro as a symbolic structure, a Place, integrating certain aspects of the Alejandro-Place with the person type to form the composite expression “deaf people.” In this construction, both the sign PERSON and a person in the discourse context, Alejandro, are symbolic components that integrate to form a composite structure. The signer can also move about, occupying different meaningful locations. In these cases the signer is a symbolic structure which can be placed, either to create a Place or to recruit an existing Place. We call constructions of this type placing the signer. We analyze the formal changes the signer makes in narrative reporting of dialogue and events – that is, changes in body orientation, eye gaze, and deixis – to be instances of placing the signer.

As always, the visual nature of signed languages influences both form and meaning. Signed languages are, quite literally, face-to-face visual languages. In signed language interaction, the canonical communication configuration is for one signer to face another signer at some culturally determined distance. Between the two is a line of sight. We call this the canonical interactional configuration (CIC). The well-known Gricean cooperative principle and maxims of conversation describe how people achieve effective conversational communication (Grice Reference Grice1989). Gricean maxims are focused on the content of the communicators’ contributions to conversational interaction. For signed communication among deaf language users, two maxims for the production of effective visual communication also come into play: (1) reduce excessive moving around from the point of view of the interlocutor, and (2) make your signs as visible as possible.

When narrating a story, canonical conversational configuration and communicative maxims constrain the signer. The signer is not free to occupy any location whatsoever, and the signs must remain visible. Thus, in Figure 22, the signer does not move into the location of the virtual interlocutor who says, “Really, I didn’t mean it.” This would result in excessive movement, especially if the narrative were to continue with a dialogue between the narrator and the virtual interlocutor: the signer would have to constantly move back and forth between these locations. It also would result in the interlocutor facing the signer, making the signs nearly impossible to be seen by the audience. Instead, what we see are attenuated changes in body orientation and eye gaze. Together, these signal that the signer occupies a new location – in our analysis, the signer has been placed.

Figure 22 was an example of reported dialogue in conversational interaction. The same placing construction is used in narrative. Figure 24 is taken from a video documenting the Argentine folk story “The Golden Hand” about three sisters who face several challenges posed by a dangerous man whose hand has been cut off and replaced with a golden hand. Prior dialogue has established a second sister Place on her right, and the Golden Hand man is on her left. The signer as narrator begins by explaining that the Golden Hand man and one of the sisters are riding together in a horse-drawn carriage. The signer’s body is in the conventional narrator Place in the center; her eye gaze and signing are oriented toward the viewing audience. In Figure 24, the narrator has been placed in the Golden Hand man Place. With his left hand the Golden Hand man keeps holding the reins of his horse; he signs with the right hand. The Golden Hand man asks: “Do you remember me? Some time ago I went to your place. There was a storm. Something happened and I lost my hand.” The narrator’s body moves to occupy a spatial position slightly on the left; eye gaze and signing are oriented toward the interlocutor, the sister Place, thus maintaining the canonical interactional configuration.

Figure 24 Golden Hand Place

(Wilcox, Martínez & Morales Reference Wilcox, Martínez, Diego, Jucker and Hausendorf2022)

In Figure 25, the narrator has been placed to occupy a spatial location on the right, the sister Place; eye gaze and signing are oriented toward the Golden Hand man Place on the right. The sister then replies: “Hmm, you? Oh, yes, I remember you came here when there was a storm. Yes, I remember your face. You cut off your hand?”

Figure 25 Sister Place

(Wilcox, Martínez & Morales Reference Wilcox, Martínez, Diego, Jucker and Hausendorf2022)

It may not seem obvious at first why we analyze this strategy for expressing reported dialogue as placing the signer: how has the signer been spatially “placed” in the same way as PERSON was placed in Figure 13 or CHANGE in Figure 14? An instructional video designed to teach students of Brazilian Sign Language (Libras) how a single signer reports interactional dialogue, such as the Golden Hand narrative, illustrates how placing the signer works. In this video the instructor, Eduardo, and his colleague, Leonardo, first demonstrate a signed dialogue as it would actually take place between two interlocutors, maintaining the face-to-face canonical interactional configuration (Figure 26).

Figure 26 Demonstration of two person interaction

(Wilcox, Martínez & Morales Reference Wilcox, Martínez, Diego, Jucker and Hausendorf2022)

Eduardo then shows how the same interaction would be presented by a single narrator with reported dialogue (Figure 27). The simplest and most realistic way for a narrator to present a two-person reported dialogue to an audience would be to “act out” the interaction by taking both roles on the phonological stage – that is, by simply recreating Figure 26, moving between the spatial locations of the two participants. This would, however, violate the visual communication maxims. Instead, Eduardo as the narrator remains essentially in one location, and by changing the orientation of his body, the narrator alternately assumes the role of Eduardo or Leonardo.

Figure 27 Narrated interaction

(Wilcox, Martínez & Morales Reference Wilcox, Martínez, Diego, Jucker and Hausendorf2022)

A diagrammatic representation of the real two-person interaction and the strategy in which the signer is placed is depicted in Figure 28. The top portion (A) shows the original interaction with Eduardo (E) on the right and Leonardo (L) on the left. The construction in which the signer is placed to express the reported interaction is depicted in the lower portion (B). When Eduardo as narrator (N) presents Eduardo’s utterances in interaction with Leonardo, he rotates his orientation slightly to his right, indicating that he has assumed Eduardo’s Place. Virtually present Leonardo assumes a position directly in front of Eduardo to maintain the canonical interactional configuration. When presenting Leonardo’s utterances directed to Eduardo, Eduardo as narrator changes orientation in the opposite direction, thus indicating that the narrator has occupied the Leonardo Place; in doing so, he takes the role of Leonardo. Eduardo, as a virtual addressee, now assumes a position in front of Leonardo to maintain the canonical interactional configuration (CIC). Thus, the overall scene is presented with Eduardo “playing” himself and Leonardo, who are both alternatively presented as virtual versions of themselves (represented by the dashed circles).

Figure 28 Placing the signer in reported dialog

(Wilcox, Martínez & Morales Reference Wilcox, Martínez, Diego, Jucker and Hausendorf2022)

In this instructional video we see the relation between a real two-person interaction and how a single narrator reports the interaction, maintaining the canonical interactional configuration while also abiding by the visual communication maxims. As it appears in the Golden Hand narrative, the placing construction closely aligns the phonological pole – the location – of the narrator with the phonological pole of the Golden Hand man. This conceptually maps the semantic pole of the narrator onto the semantic pole of the Golden Hand man. In nontechnical terms, we would say that the narrator “has become the Golden Hand” or “Eduardo has become Leonardo.”

Constructions in which the signer is placed are used in narratives to indicate changes in character perspective. By placing the signer in the Place conventionally associated with the narrator, for example, the signer is conceptually mapped to the narrator. The same process occurs for mapping the narrator to character roles, as with the Golden Hand and Libras tutorial examples. Placing the signer appears in several other grammatical constructions across different signed languages, including fictive discourse and fictive interaction (Jarque & Pascual Reference Jarque, Pascual, Dancygier, Lu and Verhagen2016); passive constructions in ASL (Janzen et al. Reference Janzen, O’Dea and Shaffer2001); and evidentiality (Jarque & Pascual Reference Jarque and Pascual2015; Shaffer Reference Shaffer, Dancygier and Sweetser2012; Wilcox & Shaffer Reference Wilcox, Shaffer and Alexandra2017).

There is a great deal more semantic and phonological complexity in constructions of this type than we have described here. It is obvious from the figures in this section that the phonological expression of placing differs. When presenting reported or constructed dialogue in conversation as in Figure 22, we see very little change in spatial location; the markers of placing are primarily the change in body orientation and in gaze direction. In the tutorial shown in Figure 27, Eduardo demonstrates how to express reported dialog by change in body orientation and eye gaze, and also with more pronounced change in upper body positioning or leaning – thus the upper body changes location somewhat more than in conversational reporting. The Golden Hand folk story narrative is presented in a more dramatic style, and so spatial placing of the narrator is more pronounced. Other factors also come into play in various types of placing the signer constructions. One complicating factor is body partitioning (Dudis Reference Dudis2004). In body partitioning, different portions of the signer’s body – hands, arms, head, mouth, eyes – may be used independently to represent different semantic content. Thus, it is possible for the signer’s body and hands to occupy a Place phonologically associated with one particular conception, while the signer’s head, eyes, and mouth occupy a different Place – looking in a different location than the one in which the body is oriented, for example – associated with a different conception. A body-partitioned double placing construction such as this might be used to simultaneously express a statement (body and hands) and the reaction to that statement by an unspecified or fictive other person (head, eyes, and mouth). It could also be used to express a statement by the signer and the signer’s stance toward that statement. For example, the signer might state that deaf people find it difficult to be accepted into medical school in some countries; while signing this, the signer turns her head and looks to her right to express a fictive interaction in which her face displays her disdain for such a situation. Here, then, the signer has used body partitioning to place her body and hands in one Place that is conceptually mapped to the signer as reporter of the factual situation, and, by shifting her head orientation and eye gaze, she creates a new Place that represents a different conceptualization – her stance in regard to the situation.

Across these examples, we see that placing constructions express conceptual mapping. Placing brings into congruence the phonological pole of a symbolic structure, whether it is a sign or the signer, with the phonological pole of another symbolic structure, thereby putting the semantic poles of these structures in some type of correspondence. In reported dialog, the correspondence is that of referential identity: by placing the signer in the phonological location of the interlocutor in a reported dialog, the signer is conceptually mapped to the interlocutor. Placing is a manifestation in actual space of the conceptual metaphors SIMILARITY IS PROXIMITY and RELATEDNESS IS PROXIMITY, which we suggest is the conceptual basis of Behaghel’s law discussed in Section 4.1. The spatial environment in which signed discourse is expressed is, in a real sense, a phonological stage on which placing constructions play out. While conceptual proximity certainly pertains to signed languages, proximity of spatial locations is an aspect of the phonology of signed languages.

6.1 From Engaged to Disengaged Cognition

We started this discussion of facial displays by showing that brow furrow and horseshoe mouth are present as behaviors outside of any linguistic system. Brow furrow is associated with physical and mental exertion. The function of horseshoe mouth is less well documented, but it appears to pertain to different states of mind, initially in reaction to the physical environment. In other words, these facial displays originate in our physically embodied interactions with the world – they are physiological responses to the perceptible environment and expressions of the two fundamental imperatives of understanding and controlling the world we inhabit. The restriction that we only have direct access to the world through perceptual and motor systems, and that this experiential interaction is the origin of embodied cognition, were two of our fundamental principles. As described by Langacker (Reference Langacker2008, pp. 535–536), “At the most basic level, we interact with the world through our senses and physical actions. There are other levels, of course: much of the world we live in is mentally and socially constructed. But either directly or indirectly, the world we construct and apprehend is grounded in sensory and motor experience.”

Direct access to the experiential world is the basis of engaged cognition. In effective control we engage the world directly when we use perceptual systems to capture perceptible targets and motor systems to move about and manipulate objects. These activities require actual physical exertion. Similarly, we exhibit epistemic control as we perceive, move around in, and attempt to make sense of the physical world. These interactions with the world require physical and mental effort and are the embodied source of exertion, reflected in brow furrow; striving to make sense of the experienced world – a world that perturbs or disturbs us in ways we attempt to make sense of and control – is the embodied source of horseshoe mouth.

Not all of our experience involves direct engagement with the world. As we develop the world we live in becomes increasingly socially, mentally, and linguistically constructed – disengaged from immediate physical experience. In other words, certain aspects of our cognition are disengaged from the world. Epistemic control is directed at mental constructions – reality conceptions of events in the world, and future or imagined realities.

A significant characteristic of disengaged cognition is that it occurs whenever engaged cognition occurs, and it recruits the same neurological and action responses that are present in engaged cognition. In order to understand this relation between engaged and disengaged cognition we must introduce the concept of subjectification. As defined in CG, subjectification refers to the process by which “mental operations inherent in experiences of a certain kind are used in abstraction from their content and applied to other circumstances (Langacker Reference Langacker2008, p. 537). An example of subjectification is the mental scanning that occurs in fictive motion (Matlock Reference Matlock2005). In conceptualizing the motion verb run when it is used to express an actual motion event, such as “The baseball player ran from first to second base” we use mental scanning to track the motion of the player along a spatial path. In using run to express fictive motion, as in “The cut ran from his wrist to his elbow,” mental scanning is directed at a static path, building up a holistic conception of the spatial configuration of the cut.

Both of these situations recruit visual perception and motor abilities: the conceptualizer visually tracks the motion of the baseball player, requiring at least a shift in eye gaze and likely movement of the head; these same physical abilities (although perhaps the movements of the eyes and head are attenuated) are used to mentally scan the cut on the man’s forearm. As we experience scenes such as this more and more, and in different contexts, we abstract a scanning scenario schema (Langacker Reference Langacker2008). The scanning involved in the scanning scenario is now fictive or simulated – no perceptual or motor actions are recruited, and the conceptualization requires only mental scanning. In saying “Basketball players are usually tall” the conceptualizer recruits her memory to remember the basketball players from her past experience, mentally scans their height, and sums up over these instances to ascribe the property of being tall as prototypical of the group. In this case, scanning is entirely a cognitive operation: no actual basketball players are visually perceived and no entity moves along a spatial path; the only “path” is the conceptualizer’s remembered experiences.

Scanning objective scenes with perceptual and motor abilities to arrive at a conceptualization engages the objective, experiential world. Scanning that is only mental, with attenuated or no direct sensory or motor engagement – as in the scanning operation used to conceptualize fictive motion and the basketball players scanning scenario – is disengaged from the experiential world. Although sensory and motor abilities are not involved, the conceptual operations that are inherent in experiencing the world are recruited, but now they are used in abstraction from any experiential content. This is the disengaged cognition used when saying usually to conceptualize an abstract scene.

How does subjectification pertain to engaged and disengaged cognition? Figure 43 (Langacker Reference Langacker2008) shows the relation between engaged and disengaged cognition. In engaged cognition (a), we interact directly with the world (W) by means of embodied perceptual and motor systems (double arrow). Box A represents engaged cognition based on direct physical experience. Diagram (b) depicts disengaged cognition, where certain aspects of engaged cognition (A) are used autonomously (A’), without any direct interaction with the world. These disengaged elements of engaged cognition are simulations of A. As we have seen, aspects of cognition that are used without any direct interaction with the world are subjectified: subjectification is one way in which cognition is disengaged from physical, perceptual, motor experience.

Figure 43 Engaged and disengaged cognition

(Langacker Reference Langacker2008 © Oxford University Press)

Facial displays are transformed from cognitively and experientially engaged to disengaged by subjectification. For brow furrow and exertion, the subjectification of engaged cognition begins in physical activity, where effective control is objective, embodied, and requires effort. The cognitive processes required to support such activity are engaged with the world. In linguistic effective control, exertion is still objective, being directed at an interlocutor and aimed at influencing reality. When a signer asks a content question, the interlocutor is expected to produce a response. Here, however, effective control is directed at the social world and not the physical world. No actual physical force is exerted in questions, assertions, or even in ordering someone to do something; the exertion here is simulated. Epistemic control requires effort to be expended, but it is the subjectified mental exertion of disengaged cognition. A similar path can be posited for epistemic control. Here, the baseline level of making sense is contact with the perceptible, physical world. This is engaged experience, and the stage at which horseshoe mouth appears, for example in infants (and even fetuses). Disengaged, linguistic epistemic control is not directed at having an effect on the world, but at creating and updating a conception of the world. It is, we would say, the ongoing process of making sense.

In all the cases of brow furrow and horseshoe mouth discussed here, these facial displays are associated with disengaged cognition and are simulations of their engaged counterparts. While the specific grammatical meanings expressed by these facial displays undoubtedly vary according to the signed language and the constructions in which they appear, the fact that these facial displays express such similar functions across unrelated signed languages is perhaps a reflection of their universal biological heritage that originates in our cognitively engaged, physically embodied interactions with the world.

7.1 Sign Phonology Redux

We learned in Section 1 that Stokoe’s pioneering research set in motion the modern linguistic study of signed languages. His body of work from the late 1950s to the 1990s spanned the history of linguistics from structuralism to the emergence of cognitive linguistics. While cognitive linguistics has now deeply influenced the field of sign linguistics, approaches to sign phonology have remained rooted in structuralism. Structuralist assumptions about the nature of sign phonology have created profound problems for the description and analysis of signed language grammar, raising such issues as the highly iconic nature of signs, the search for signed (visual, optical) equivalents to segments, syllables, prosody and stress, and the relation between phonology and phonetics. The most surprising trend has been the re-emergence of gesture in the description of the grammars of signed languages with the claim that many signs and sign constructions are fusions of language and gesture.

Although Stokoe was not a linguist by training, his academic background, which culminated in a PhD in English specializing in medieval literature, provided him with extensive knowledge of how languages work and how they change. He wrote convincingly, for example, of phonology, morphology, and syntax in Middle English (Stokoe Reference Stokoe1955), as well as broader issues of language use, language change, and the interaction of language and culture. Looking back at this experience, he wrote (Stokoe Reference Stokoe1994, pp. 377–378):

When he began teaching deaf students at Gallaudet College, he recognized that their signing was a language. When he was told by his colleagues it was just broken English and gesture he knew something was amiss, and thus started his journey to prove what he already knew: his deaf students were using a language. He turned to the predominant linguists and linguistic theory of the day, studying with linguists George Trager and Henry Lee Smith at the 1957 Summer Institute of Linguistics. Trager and Smith were American structuralists, and thus Stokoe’s first analysis was steeped in structuralist linguistic theory.

As we saw in Section 1, Stokoe regarded cheremes as the structural equivalent of phonemes, writing that “Like consonant and vowel,” the cheremes of position, configuration, and motion “may only be described in terms of contrast with each other” (Stokoe Reference Stokoe2005, p. 20); in other words, the mode of analysis was based on identifying distinctive features. Even at the early stage of his work, however, Stokoe was keenly aware of the important differences between spoken and signed language, especially at the level of phonology. He began to resist the constraints of structuralism and application of theory derived from spoken language. Writing later of this period (Stokoe Reference Stokoe1980b), he reported recognizing what he regarded as a central difference: while the phonological elements in spoken languages can only be arranged linearly in the dimension of time (except for stress and intonation), signed languages allow and exploit the three dimensions of space and one of time. His revised thinking on sign structure is worth citing in full (Stokoe Reference Stokoe1980b, p. 369):

By 1980, the influential The Signs of Language (Klima & Bellugi Reference Klima and Bellugi1979) had appeared. The term parameter was now being used and understood by most sign linguists as synonymous with Stokoe’s term aspect. He saw this as a serious error, writing, “More often in recent research the term parameter is used for what is here called an aspect. There is more in this than a quibble over words, which a nonlinguist might suspect” (Stokoe Reference Stokoe1980b, p. 371). Stokoe further clarified his view of sign phonology. First, he explained why he considered the switch in terminology from aspects to parameters to be more than a quibble: “My own preference, more philological than linguistic, would be to leave ‘parameters’ to those who work with quantities and measures generally accepted as quantifiable, and to keep ‘aspects’ in its literal sense of ‘what appears to a specific way of looking’” (Stokoe Reference Stokoe1980a, p. 895). A key word here is “quantifiable.” In Section 7.2 we will see problems of quantifiability, listability, and countability return. What Stokoe saw as more than a quibble has turned in recent years into a full-fledged disagreement over what distinguishes – and the irony is inescapable – sign from gesture.

Stokoe also observed that another problem arises when the aspect of location is renamed “place of articulation” and considered to be analogous to the concept from spoken language phonetics. Location in signed language, he noted, is not equivalent to place of articulation in speech. The difference is once again because signs occur in visible space. “What is seen is the sign, whereas what the organs of articulation do is shape the sound that is heard. The place or location is always of significance in signing; a sign is where it is” (Stokoe Reference Stokoe1980a, p. 896).

Stokoe continued to advance his ideas about sign phonology, moving even farther from structuralism in an article entitled “Semantic Phonology” (Stokoe Reference Stokoe1991). Influenced by Wallace Chafe’s Meaning and the Structure of Language (Chafe Reference Chafe1970), Stokoe reiterated his view that in analyzing the form of a sign, one needs only to think of a sign as something that acts together with its action. From there, he made the conceptual leap of giving meaning to form itself (Stokoe Reference Stokoe1991, p. 107):

In a sense, semantic phonology merely elaborated Stokoe’s original vision of sign phonology presented in 1960. The essential idea was there all along: “some thing” that acts, and its action, the semantic phonological noun and verb. The seed of his aspectual analysis blossomed into what might be seen as the inception of a conceptual approach to sign phonology. Using Cognitive Grammar concepts, we would say that semantic phonology views the thing that acts and its action as conceptual archetypes for prototypical nouns and verbs, as described in Section 3.2. Like prototypical nouns, the acting thing is substantive and conceptually autonomous, residing primarily in space. Its action is conceptually dependent, prototypically energetic, and resides primarily in (change over) time. As an event, action tout court is not possible – action requires some thing which performs the action. This means we can also describe semantic phonology in terms of baseline and elaboration. A sign action such as a path movement performs an operation on a more substantive baseline actor, prototypically a hand.

Although Stokoe did not have the conceptual theoretical framework on which to hang his idea of semantic phonology, he did see that it was a new way to view the relation between form and meaning, breaking down the traditional linear view of language where “first there are the sounds (phonology), these are put together to make the words and their classes (morphology), the words in turn, are found to be of various classes, and these are used to form phrase structures (grammar), and finally – the delay is built into this pedestrian way of thinking – the phrase structures after lexical replacement of their symbols yields meaning (semantics)” (Stokoe Reference Stokoe1991, p. 112). Instead, he conceived of language in which “semantic phonology ties the last step to the first, making a seamless circuit of this pitty-pat progression. The metaphor for semantic phonology that jumps to mind is the Möbius strip: the input is the output – with a twist!” Stokoe regarded semantic phonology as a more general account for another reason as well: under this conception the actor is not restricted to hands. Facial articulators may equally well serve as actors and their actions: “Semantic phonology admits all kinds and colors of sign activity, welcoming nonmanual as well as manual agents and all their actions” (Stokoe Reference Stokoe1991, p. 108). Facial actions such as those we described in Section 6 may also be seen as operations on a substantive articulatory base (raising eyebrows, furrowing the forehead, lowering the corners of the mouth, squinting the eyes).

In a sense, semantic phonology represented not only the culmination of Stokoe’s ongoing journey to move beyond structuralist notions, it also reflected his driving motivation to find simpler solutions. We noted that Stokoe was not initially a linguist. We cannot resist concluding this brief historical sketch with a nod to an article on the topic of Stokoe’s original area of expertise, medieval literature. In it, we see a preview of his style of thinking a decade before he became involved in signed languages. Critiquing another scholar’s analysis of two sources for the Middle English metrical romance Sir Launfal, he wrote, “This theory is not only more complicated than it need be to explain two uncomplicated stories, but it also disparages unjustly the story-telling ability of both the anonymous author and Marie de France” (Stokoe Reference Stokoe1948, p. 393). In one succinct sentence Stokoe shows us his preference for simplifying over complicating. Thirty-three years later, in describing semantic phonology, we see the same preference for simplifying: “one needs only to think of a sign as something that acts together with its action; it’s that simple: no features, no autosegments, no orientation, no contacting or contacted parts, no HOLDs, no MOVEs, no tiers, no tears – just something acting (limb or limbs, head, face, whatever) along with its action” (Stokoe Reference Stokoe1991, pp. 111–112).

In summary, sign phonology began firmly rooted in structuralist thinking. The structuralist approach assumes a universal inventory of phonological elements. Signed language phonological theory adopted these assumptions and searched for a manual phonological inventory of meaningless elements as well. Stokoe’s initial analysis and his later dictionary of American Sign Language (Stokoe et al. Reference Stokoe, Casterline and Croneberg1965) did, in fact, offer such an inventory, as have other sign linguists (Liddell & Johnson Reference Liddell and Johnson1989; Klima & Bellugi Reference Klima and Bellugi1979). However, decades of phonetic research on spoken language has shown that no inventory of countable phonetic objects exists (Port & Leary Reference Port and Leary2005). This countable inventory problem is most obvious in attempts to deal with phonological location. Bodily locations – nose, eyes, mouth, hands, fingers – would seem to be tractable problems, easily countable. The problem arises when referents used in grammatical constructions, such as those described in Section 4, occupy spatial locations in the spatiotemporal discourse environment. We discuss the nature of this problem and attempt to resolve it in the next section.

7.2 Location, Location, Location

The historical arc of our understanding of signed languages has taken a surprising trajectory. For centuries signs were regarded as nothing more than holistic, imagistic unanalyzable gesture. With the advent of modern linguistic approaches pioneered by Stokoe, signs were shown to be analyzable based on linguistic principles. Signed languages exhibit the same types of structures – phonological, morphological, syntactic, discourse – as spoken language. Gesture has re-entered the scene in recent years with the claim that signs are composites, or fusions, of language and gesture (Schembri et al. Reference Schembri, Cormier and Fenlon2018; Meier & Lillo-Martin Reference Meier and Lillo-Martin2013; Lillo-Martin & Meier Reference Lillo-Martin and Richard2011). For the reader interested in pursuing the various positions, Wilbur (Reference Wilbur2013) offers a comprehensive critique from a more formalist perspective. Dotter (Reference Dotter2018) gives a detailed and comprehensive summary of the assumptions made by what he calls the “Gesture School” of sign linguistics from a typological and cognitive perspective.

The first step in arguing for signs to be part language and part gesture has been to distinguish the two. A number of criteria have been offered (Brentari Reference Brentari2019; McNeill Reference McNeill1992; McNeill Reference McNeill2000; Sandler Reference Sandler2009), including:

• Gesture lacks hierarchical combinatorial structure

• Gesture is more gradient than language, insofar as the gradience in language is anchored to the language’s phonological categories, while gestures are not so constrained.

• Gesture is more variable than language. Sign languages have an established lexicon and morphological categories, while gestures do not.

• Co-speech and co-sign gestures are often produced below the level of awareness of speakers and signers, and access implicit, nondeclarative knowledge.

• In language, but not gesture, the phonological and morphological systems distribute their contrasts in a symmetrical fashion within the articulatory and perceptual space using principles of paradigm uniformity

• Gestures are holistic, noncombinatoric, idiosyncratic, and context-sensitive. The linguistic signal is the converse: dually patterned, combinatoric, conventionalized, and far less context-dependent in the relevant sense.

While these criteria may apply to all aspects of sign structure, the predominant current concern focuses on location. Much of the debate questioning whether location is language or gesture has centered around pointing, pronouns, and verb agreement. Liddell (Reference Liddell2003) provides an array of examples pointing out the difficulties with specifying locations in pointing constructions (which Liddell calls pointing gestures); reciprocal verbs; indicating verbs (verbs capable of being meaningfully directed in space toward entities, directions, or places); pronouns; and depicting verbs (verbs which include features depicting certain aspects of their meaning). The problem arises because these constructions can use innumerable locations. For example, consider the ASL verb ASK-QUESTION. When a signer asks an addressee a question, ASK-QUESTION is directed at the addressee’s chin. The location in three-dimensional space changes when the addressee is much taller than the signer. As Liddell (Reference Liddell and Lucas1990) observes, if the signer imagined facing an exceptionally tall US basketball player, and she wanted to ask him a question, ASK-QUESTION would be directed to a location considerably higher than the signer. Of course, many other factors can determine the location of the addressee. No matter where the addressee is located, ASK-QUESTION is directed at the chin.

Other signs must be directed at other specific parts of the body. GIVE, for example, is directed at the addressee’s chest, and COMMUNICATE-TELEPATHICALLY is directed at the addressee’s forehead (Liddell Reference Liddell2003). In all of these cases, where the addressee is located, or where a nonpresent addressee is conceptualized as being located, will determine where these signs are directed.

Depicting signs exhibit the same properties. The verb UPRIGHT-PERSON-WALK-ALONG is produced with a path movement. When produced with a particular facial display, the sign means “person walk along in an unhurried movement.” As Liddell describes the problem, the sign combines a lexically fixed verb (the index finger symbolizing the upright person) with “variable and gradient initial and final locations” (Liddell Reference Liddell2003, pp. 270–271). As we have seen, Liddell associates these properties with gesture.

Liddell Reference Liddell2003 makes the same case for ASL pronouns. He poses the situation in which a signer is standing on a stage in an auditorium with hundreds of seats, some lower than the stage, some at the level of the stage, and some higher in balconies. In signing YOU to any individual in the auditorium the pronoun has to be directed at the referent, who could be sitting in any of the seats. The direction of pointing – that is, the location of the referent – changes from one referent to the next. Liddell again attributes this to gradience and concludes that these facts about location “are inconsistent with the claim that there is a locus associated with the addressee toward which signs are directed. If there were such a locus, all directional signs referring to the same entity (e.g. the addressee) would be directed toward that single locus” (Reference Liddell2003, p. 76). In Liddell’s analysis these loci are gradient, variable, uncountable – and thus they are gesture.

But are the different locations toward which signs are directed properly described as gradient? An analogy will provide an answer. Suppose an Olympic archer shoots at a target 50 yards away and 6 feet off the ground. Being a skilled archer, the arrow lands in the center ring of the target. No matter how skilled she is, however, if she continues to shoot arrows at the target they will not land in precisely the same spot. They will disperse in locations around the center ring. This is gradience. The archers aim at one target, but where the multiple shots land will define an area of gradience around the center. If we set up another target and the archer shoots multiple arrows at it, the same will be true. The archer aims at two different locations. Again, multiple hits will define a gradient area around the center. Nevertheless, the archer is shooting at two locations. The fact that there are two locations does not constitute gradience. There is not one immense, gradient target, and the archer does not shoot arrows knowing that she has to hit two targets in order to win, hoping that gradience will direct them to two distinct targets. We can keep adding targets – in fact, there is no necessary end to the number of targets our archer could shoot arrows at. We will get the same result: the number of targets is not gradience.

As the analogy pertains to locations in signed languages, the targets are the final locations of the verbs GIVE, ASK-QUESTION, and COMMUNICATE-TELEPATHICALLY. Just as the precise location where multiple arrows aimed at the same target land will exhibit gradience, the signer’s index finger in the final location of these verbs will not always hit the same spatial target. But the signer is directing her hand at a specific location: the addressee’s chest, chin, or forehead, no matter where the addressee is or is conceived to be. The central problem is confusing production with conceptualization. Gradience pertains to production. The number and location of real or conceived referents is a matter of conceptualization.

These data presented here pose two theoretical issues, both pertaining to the nature of grammar and to sign phonology: (1) the number of possible locations, and (2) how the connection between a referent and its phonological location is specified. Both issues arise because of how sign linguistics approaches phonology. For many sign linguists, phonology assumes a list – a finite inventory of distinctive phonemes. The inability to list every possible phonological location leads sign linguists to claim that these locations cannot be part of a grammatical system, and therefore they must be gesture. Liddell recognizes the need to specify phonological locations, but we saw he attributes these specifications to gradience and thus gesture: “Each individual verb has specific gestural characteristics associated with it” (Reference Liddell2003, p. 139). Wilbur (Reference Wilbur2013, pp. 222–223), although she does not accept this argument, summarizes it succinctly with regard to the first problem:

The ramifications are profound because constructions of the type under discussion are pervasive in signed languages. By this reasoning, these types of constructions must be fusions of gestural locations and linguistic elements (Schembri et al. Reference Schembri, Cormier and Fenlon2018), or blends “in which gestural and linguistic elements are co-produced within a single sign” (Özyürek Reference Asli., Pfau, Steinbach and Woll2012, p. 637).

Concerning the second theoretical issue, Lillo-Martin and Meier (Reference Lillo-Martin and Richard2011, p. 121) adopt a theory in which “actual real-world locations of referents are not part of the grammar, so in order for a linguistic object to ‘point to’ such locations, language must interface closely with the gestural system.” In Lillo-Martin and Meier’s approach, abstract indices are defined in the grammar, whereas real-world loci are outside of grammar. This constrains how they can explain the connection between a referent and its location. As a result, they conclude that “the grammar doesn’t care which point in space is used for a particular referent … the connection between referents and loci requires language to interface with gesture” (Lillo-Martin & Meier Reference Lillo-Martin and Richard2011, p. 121). Liddell relates this problem to phonology, saying that “describing a spatial locus in terms of some type of phonological or phonetic feature has proven to be extremely difficult and there is currently no satisfactory feature system capable of doing this” (Liddell Reference Liddell2003, p. 136).

Such theoretical approaches lead to the untenable position that a core element of signed languages is gesture. The grammar of a signed language is incapable of telling signers how to point to a real-world location, such as a cup of cappuccino, and say “This is the best cappuccino I’ve ever had.” Surprisingly, gesture is able to do all of this quite well, so well that even a child can do it. One might ask: how is gesture able to do what grammar cannot? Does gesture keep a list of all the locations that a person can point to? Does gesture keep a list of all the locations of addressee’s chins, whether the addressee was taller or shorter than the signer, lying on the floor, standing on the roof, or rising aloft in a hot air balloon? We suspect that is not how gesture handles what grammar cannot. But the suggestion that grammar has to interface with gesture clearly recognizes that a solution to the problem does exist – it’s just that this approach to grammar cannot solve it, and so the problem is handed off to gesture. We might ask why, once we understand how gesture handles the problem, grammar could not use the same solution, unless linguists have so handcuffed grammar that we are left to grasp at straws in order to conform to self-imposed theoretical constraints.

Traditional linguistic theory presupposes that semantics and phonology – the meaningful and the meaningless – are independent domains of language. As a result of this disjunction, the ability of the grammar to connect referents with their locations breaks down and the problem must be turned over to gesture. It should come as no surprise that we do not accept a view of language that requires two distinct fields of cognitive potential and an interface between the two. When a signer places and points to two locations on her left and right to compare two abstract ideas, perhaps pilot wave theory and many-worlds theory alternatives to the Copenhagen interpretation of quantum mechanics; when a signer directs signs at a location on her left when telling a story about a conversation with an imaginary Golden Hand man; when a signer points to the former location of a building that was torn down and is no longer there; when a mother directs the sign FINISH at a lower location when telling her deaf son to stop misbehaving; when a young deaf boy directs the ASL sign ASK upward when talking to his favorite basketball player; when a signer points to and directs her eye gaze at the addressee standing in front of her; when a signer points to the location of an absent cappuccino cup that only a few seconds previously was on the table and asks where it went; when a signer points to her iPhone and signs, “that’s my iPhone”; when a signer points to her nose to mean “nose” – all of these usage events make use of symbolic structures, an association of a semantic pole and a phonological pole, symbolic structures that we have analyzed as Places. In all cases, the conceptualized location of an entity is symbolically realized as an actual phonological location. We see no reason why some of these locations should be accepted as linguistic and others denied linguistic citizenship.

Places are symbolic structures consisting of a schematic semantic pole (“thing”) and a schematic phonological pole (a location in space). Through our experience with the world we construct intersecting hierarchical networks of Places with varying degrees of semantic and phonological schematicity/specificity. Cognitive abilities such as memory and imagination operate to yield other types of Places, with increased detachment from perception of the experiential world requiring increased conceptual resources. By using these general cognitive abilities and the linguistic resource of Places, signers are able to instantiate and direct attention to an unlimited number of Places that refer to distinct discourse entities.

This approach may not seem like the type of simple, parsimonious model of grammar that a linguist should propose – better to deport the uncountable offending elements. Our reply would be, “we judge theories not by the number of entities they can and do describe but by the simplicity of their underlying ideas.”Footnote ⁹ Our approach has a single underlying idea. Our account is grounded in established Cognitive Grammar and does not postulate new theoretical constructs. Places are simply symbolic structures that exploit space for their phonological expression. Yes, we posit a conceptual network of Places, and we recruit cognitive abilities. But linguistic theories already have semantic and phonological networks, and the cognitive abilities are already well established, for both linguistic and nonlinguistic conceptualization. On the other hand, approaches that require the most common elements of signed language expressions to be fusions of two distinct systems, and suggest that knowing how to point to and talk about an object in the environment requires a system distinct from language, with the additional complication of an interface between the two – these approaches seem not so simple.

Liddell (Reference Liddell2003) develops an extensive cognitive account of ASL grammar based on mental space and blending theory (Fauconnier Reference Fauconnier1985; Fauconnier Reference Fauconnier1997; Fauconnier & Turner Reference Fauconnier, Turner and Goldberg1996). Mental space theory is based on cognitive linguistic principles, including the experiential grounding of language and cognition in human’s perceptual-motor interaction with the spatial environment. And yet, Liddell’s model imposes the structural barrier between phonological space and semantic space that forces him to classify certain locations as gestural. His 2003 account holds to the assumption that ASL “consists of more than conventional linguistic forms. It also includes gradient aspects of the signal (typically directional aspects or placement) and gestures of various types.” As a result, Liddell believed that “ignoring gestures and gradient aspects of the signal produced by the hands leaves out too much” and that “restricting the analysis of language to symbolic units and their grammatical organization is hopelessly inadequate” (2003, p. 362). The problem is these are not gradient, nonconventional, nonsymbolic aspects of ASL. Constructions that incorporate the spatial locations necessarily used in pointing, placing, and multiple symbolization are fully analyzable as conventional, symbolic linguistic units.

Finally, we should note that in their latest work, Johnson and Liddell (Reference Liddell and Johnson2019, Reference Johnson and Liddell2021) do not use the term gesture. Rather, they propose a phonetic system based on placement bearings – the term “bearing” borrowed from its navigational meaning – which describe the direction a hand points. They also divide signed languages into two broad categories of signs: indicating signs (agreement, directional, etc.) such as SAY-NO-TO→y and ASK-QUESTION, and nonindicating signs such as MOTHER, EAT, TREE, LIKE.

The clash of a cognitive linguistic approach to the grammar of signed languages as developed by Liddell and a lingering structural approach to sign phonology plays out in this proposed phonetic system. In indicating signs, the bearings are “ad hoc pointing directions the signer must conceptualize and receivers must construct in their own minds” (Liddell & Johnson Reference Liddell and Johnson2019, p. 167). In contrast, nonindicating signs require a “phonetic inventory of lexically fixed bearings.” These two contrasting requirements derive from the same problem caused by the uncountable spatial locations of referents, particularly in indicating signs. As we have seen, these spatial locations were branded as gesture and excluded from the linguistic system. The current proposal is simply silent on spatial locations: “spatial loci no longer play a role in our phonetic representations of hand placement” (ibid., p. 167).

Navigational systems certainly work. Clearly, however, navigational systems imply a destination: in a nautical navigational system the destination is the spatial location toward which the ship sails. Ships do not set a bearing simply for the sake of setting a bearing. The raison d’être of the bearing is to reach the destination. For signers, the sole purpose of a hand placement bearing for indicating signs is to identify for the interlocutor a referent occupying a particular location in space. Johnson and Liddell acknowledge, as they must, not only that the spatial locations are the destinations of the hand bearing, but that they remain (at least in their view) uncountable: “indicating signs require a different conception of phonetics, one that includes nonlistable phonetic features (i.e., placement bearings and displacement bearings that point at people and things)” (ibid., p. 167).

The move to a hand placement bearing approach strikes us as sleight-of-hand, and yet another complication imposed by structuralist thinking: rather than admitting spatial locations into the grammar of language and describing them directly, the proposed model describes the bearings that indirectly point to entities which signers know to occupy spatial locations. As Johnson and Liddell state (ibid., p. 145): “It is important to note that we are proposing an uncountable number of ways an indicating verb can point, not an uncountable number of meaningful spatial loci or an uncountable number of grammatical elements.”

Signed languages use space. Signers use locations in space for the purpose of constructing meaning. If existing linguistic theories cannot account for such things, then surely this is too bad for linguistic theories. Turn the page and start rethinking. Our proposal, we believe, is at once simpler and targets what the signer is actually conceptualizing. Meaningful spatial locations are elements of the expression of signed languages – they are the phonological pole of Place symbolic structures. This proposal does not require an uncountable number of grammatical elements. It needs only one, the grammatical element of Place, a schematic symbolic structure which, like the grammatical element noun, has uncountable instantiations.

Cognitive Grammar holds that phonological space is a subregion of semantic space, and thus a linguistic symbol is “a correspondence between two structures in semantic space (broadly conceived) where one occupies the phonological subregion” (Langacker Reference Langacker1987, p. 79). In other words, phonological structures are conceptual. The phonological location of Place is a conceptual entity. How phonological entities are realized phonetically certainly requires explication. We believe the best solution will lie in task dynamic approaches such as has been applied to speech and to motor action (Browman & Goldstein Reference Browman and Goldstein2010b; Hawkins Reference Hawkins, Docherty and Ladd1992; Saltzman & Kelso Reference Saltzman and Scott Kelso2009; Thelen & Smith Reference Thelen and Linda1994). One such model is articulatory phonology (Browman & Goldstein Reference Browman and Goldstein1990; Browman & Goldstein Reference Browman and Goldstein2010a). Application of such models to signed language would seem a natural extension, but little work has been done toward making the connection (see, however, Tyrone et al. Reference Tyrone, Nam, Saltzman, Mathur and Goldstein2010 and Silva & Xavier Reference Silva and Nogueira Xavier2018).

7.3 Science or Ethnoscience?

This leaves us with the question: what is gesture? The urge to distinguish between language and gesture is ancient. For most of this time the distinction appeared simple and natural: language is speech, gesture is predominantly what the hands do, including signing. This simple solution of equating language with speech and gesture with the hands does not work in the face of signed languages. As we have just seen, sign linguists began to look for approaches to distinguish language and gesture based on more refined criteria; some called for defining a modality free notion of gesture (Okrent Reference Okrent, Meier, Cormier and Quinto-Pozos2002), with little substance on what such a notion would entail.

We question whether objective criteria can ever be specified that will uniquely distinguish “language” from “gesture.” In light of what we now know about signed languages and the multimodality of communicative behavior, we question whether “gesture” is even a useful scientific concept. Adam Kendon, one of the most distinguished scholars working on gesture today, raises the same question:

Kendon argues that it is time to discard the categories “gesture” and “sign” and develop a comparative semiotics of visible bodily action (kinesis) used in utterances by speakers and by signers. He suggests that doing so would “resolve and clarify the otherwise rather fraught discussions of how ‘gesture’ and sign’ are related, as well as the problems encountered when, in a signed utterance, we sometimes have difficulty in deciding whether a given expression is a ‘gesture’ or a ‘sign’” (Kendon Reference Kendon2017, p. 30).

We use “perceptible usage events” as a term that includes bodily actions that are visually perceived as well as bodily actions that produce an acoustic signal that is perceived by the auditory system. These are the same bodily actions that, as we discussed in Section 2, Neisser (Reference Neisser1967) calls articulatory gestures, and the gestural units of articulatory phonology described at the end of the previous section. In this sense, spoken language, signed language, and gesture are all gestural. As we have noted, perceptible events do not come to language users as sorted, listed, and labeled phenomena. The search for categorical, objective, observable criteria that will distinguish between language and gesture seems to us the wrong approach. “Language” and “gesture” are not, we submit, scientific concepts. They are ethnoscientific concepts.

We are reminded of two stories told by cultural anthropologists. The first is from William Sturtevant (Reference Sturtevant1964, p. 100), who defines culture as

The second story comes from the cultural anthropologist Clifford Geertz (Reference Geertz1973, p. 6):

The answer to figuring out what is what lies with the observer, not the observed. As early as Reference Stokoe1960, Stokoe recognized that sign linguists should not impose criteria to distinguish sign from gesture; rather, they should see the world as a deaf person would. As he wrote (Stokoe Reference Stokoe1960, pp. 6–7):

Stokoe’s shrug, Ryle’s wink, and the other perceptible events deaf people experience every moment of their life is the stuff from which language is made. The same behavior, with a different fleck of experience and a speck of categorization by the observer, and – voilà! – language. The answer to Kendon’s question of whether a given expression is a “gesture” or a “sign,” or to whether certain expressions are language-gesture fusions or not, lies not in finding observable, linguist-as-a-camera photographable differences between “sign” and “gesture.” It is a matter of what counts as sign or gesture from the observer’s point of view. This, in turn, depends on received folk classifications that are handed down and change over time, that vary across cultures and across the individuals doing the classification. “Language,” “gesture,” and “sign” are historical-cultural constructs, folk classifications. Whether deaf and hearing people, not to mention regular folk and linguists, share the same classifications of what counts as sign and what counts as gesture – if indeed communities of signed language users even have such categories – is an open and important question. Research suggests that the answer is likely to be quite complex (Kusters & Sahasrabudhe Reference Kusters and Sahasrabudhe2018).