3.1. grammaticalization analyses

Many linguists working in grammaticalization theory have discussed be going to (e.g., Bybee et al., Reference Bybee, Perkins and Pagliuca1994; Hopper & Traugott, Reference Hopper and Traugott1993; Traugott & Dasher, Reference Traugott and Dasher2002). These analyses have primarily limited themselves to explaining the shift from be going to solely indicating motion through space to its development as a future marker. Like Langacker (Reference Langacker1987) and Brisard (Reference Brisard2001), these researchers also reject the adequacy of the general metaphor time is motion to explain the development of the future readings. However, all the grammaticalization analyses agree with Sweetser’s (Reference Sweetser1990) claim that semantic extension tends to follow the path of a lexical unit first having a spatial/physical meaning from which a more subjective, abstract meaning develops; Sweetser discusses these extensions in rather general terms of metaphorical extension.

Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) undertook a large-scale, cross-linguistic corpus study of lexical items that equate with come/go and develop a future marker interpretation. They discovered that the future reading only arises in the context of a construction that also contains a marker of imperfective aspect and an allative marker (indicating that the motion is toward a goal). Traugott and Dasher (Reference Traugott and Dasher2002) discuss this finding and reiterate Bybee et al.’s (Reference Bybee, Perkins and Pagliuca1994) interpretation that the future meaning seems to arise from invited inferences of temporality and intentionality inherent in the construction:

Note that a central part of the argument is that movement through space, especially forward movement along a path towards a goal, inherently involves a temporal element or invites an inherent inference of temporality. According to Traugott and Dasher (Reference Traugott and Dasher2002), the presence of the imperfective aspect reinforces the notion that the spatial process is taking place in time: “motion towards something takes time (i.e., is imperfective) and one will arrive there only at some time later than the motion starts” (p. 83). They argue that, over many uses, “the motion through space meaning can be backgrounded while the inherent temporal meaning is highlighted and through pragmatic strengthening can come to be independently associated with the construction as a distinct meaning” (pp. 83–84). Importantly, Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) further argue that the invited inference of intentionality is also a key foundational element for the future interpretation. Traugott and Dasher offer historical evidence from written text that be going to was first used primarily to indicate motion through space and only later also came to be used to indicate temporality (i.e., future).

We find several important insights from the research in grammaticalization: (1) across many languages, patterns of extension are systemic and motivated, going from the spatial/physical to the more subjective and abstract; (2) semantic extensions can be based in invited inferences that are present when a lexical unit is used to indicate its core spatial/physical meaning; (3) semantic extension is gradual, and can go from nuanced highlighting of a component of the lexical unit’s meaning to an independent, distinct sense; (4) semantic extension results in polysemy networks in which the original spatial/physical meaning is the core and often continues as one of the primary uses of the lexical item (note that be going to continues to indicate physical/spatial movement towards a goal, as in I’m going to the store); and (5) the usage-event of declaring one is about to do something entails the announcement of the intention to do that thing. Nevertheless, we argue that the explanations as such are slightly askew and do not adequately explain either be going to’s extension from the spatial/physical motion interpretation to the neutral future interpretation. We will argue that several elements of embodied semantics are missing from these earlier accounts. Particularly important is the fact that the analysis does not account for the several additional senses associated with be going to. Moreover, these analyses do not address the semantic relationship between will and be going to. We will argue that the analysis does not sufficiently take into consideration the etymological roots of be going to and will; close examination of earlier attested meanings reveals important traces of the original meaning, which continue to constrain and color the extended, present uses. Evans (Reference Evans2009, Reference Evans, Evans and Chilton2010, Reference Evans2014, Reference Evans2015b) discusses this phenomenon in terms of semantic parameters and inter-lexical polysemy in his analysis of the subtle, but important differences between the state sense associated with English in and the state sense associated with English on.

3.2. will

Historically, the first attested uses of the cognates of will can be traced as far back as Sanskrit vrnoti, which indicated choosing or strongly desiring. Although the form has undergone considerable change, strong reflexes of the original meaning of strong desire or wish have persisted. According to the Online Etymological Dictionary, the reconstructed Proto-Indo-European form *wel meant ‘to wish’, based on cognates in Old Frisian, Old Norse, Old Saxon, and Old English which all have verb cognates that meant ‘to desire, to wish’. Interestingly, the Old English cognate wel was already developing a future use. Moreover, wel included an implication of intention or volition which distinguished it from sheele ‘shall’, which indicated obligation. ^{Footnote 2}

According to the OED, the earlier cognate forms of desiring or wishing were often expressed by an authority figure, so that a desired act or event was considered compelled. Thus, the strong desire of a god or person of authority caused the state or act to be accomplished, as in:

1. (12) The king wills these laws be enforced.

In other words, it is the king’s desire, born of his own internal volition, that the laws be enforced. Further, in an interactive, social situation, an authority figure’s announcement of a strong desire to have something accomplished is likely to evoke a response by the addressee of expressing the volition to see that the authority figure’s desire is fulfilled. Thus, many instances in which will would have occurred not only expressed strong desire but also strong volition. In either case, i.e., the authority figure announcing a strong desire or the addressee announcing a strong volition to fulfill that desire, the initiating forces in these historical uses of will are internal to particular actors. Through grammaticalization, certain uses of will have been bleached of this internal, volitional force (Traugott & Dasher, Reference Traugott and Dasher2002). Traugott and Dasher provide evidence that by Middle English will had developed the neutral, scheduled future sense. Tyler’s (Reference Tyler, Robinson and Ellis2008) analysis argues that this neutral future sense initially stems from the speaker’s certainty about her own desire and volition to fulfill some future goal.

Tyler (Reference Tyler, Robinson and Ellis2008) did not develop the link between the speaker’s certainty about her own desires and the actual fulfillment of those desires. Building on this earlier work, we add another element of embodied experience to the evolution of the modals. If one regularly makes a commitment and successfully fulfills it, the association between making the commitment and surety of the not-yet-realized event actually happening becomes strongly linked. Similarly, if one regularly experiences an outside authority (like the national train service or your favorite movie theater) or a reliable actor announcing a scheduled but not-yet-realized event and that event actually occurring at the announced time, one forms a strong association between announced, scheduled events and the actual fulfillment of those events. Note that the announcement of one’s strong desire or volition invites an inference that the speaker has the intention of fulfilling that desire. This argument draws on the Bybee at al. (1994) argument that the announcement of going somewhere entails an announcement of the intent to go somewhere. We would add that the announcement of strong volition not only invites an inference of strong intention, but also an inference of prediction that the not-yet-realized situation will be realized at some point in the future. Thus, the embodied experiential links between the announced desire/volition and the concurrent implied intention to fulfill the desire, on the one hand, and the realization of the event, on the other hand, become so strongly entrenched that the form will, which originally only indicated an individual actor’s strong desire or volition, came to have the additional meaning of ‘neutral’ or ‘scheduled’, ‘certain’ future.

The inferences of intention and prediction associated with the use of will are very similar to those Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) and Traugott and Dasher (Reference Traugott and Dasher2002) hypothesize for be going to. However, it would seem that such invited inferences are inherent in any usage event of communicating that one is about to do something, not only from the particular verb constructions be going to and will. When a speaker tells an addressee that she is preparing to drive the car to the university, this is a declaration of intention and a prediction the intention will be fulfilled. It is worth noting that the kind of linking between the invited inferences inherent in the use of a lexical construction (as in I’m going to the store) and the embodied experience of regularly having the entailed predictions fulfilled is not the same type of experiential linking discussed by Grady (e.g., 1997) and Lakoff and Johnson (Reference Lakoff and Johnson1996) as experiential correlation or primary metaphor. Experiential correlation refers to two independent physical/spatial events or states co-occurring so frequently that they become linked in memory, such as the independent events of adding more of a liquid to a container and the observation of a rise in vertical elevation in the level of the liquid. Once the two states or events are strongly linked in memory, language associated with one physical/spatial sense can be used to refer to the other physical/spatial scene or event. Thus, English speakers use language concerning vertical elevation to refer to the independent domain of amount, as in the well-known primary metaphor more is up. Rather, the linking between invited inferences inherent in a language usage event (i.e., particular language used in a particular communicative context) and frequent, recurring human experience, such as having the intention implied in the usage event fulfilled, seems to more akin to what Evans (Reference Evans2009, Reference Evans2014, Reference Evans2015b) discusses as the link between a lexical concept and a cognitive model. In this case, humans develop a cognitive model of speech events in which a speaker’s declaration of an intention to do something or go somewhere includes that declaration as being understood as a prediction on the part of the speaker that the intention is likely to be fulfilled. The cognitive model is accessed or activated via the language the speaker uses to declare the intended action. This is an important distinction to make as it emphasizes that meaning extension takes place by multiple routes, invited inferences and primary metaphor being two. ^{Footnote 3}

Although grammaticalization and bleaching have occurred with will, and its original meaning of volition or desire has been bleached from one important use, the neutral or scheduled use, we have already seen in other instances that the volitional sense is still active. Additionally, in current English, related forms such as willing, willingness, willfulness, and the noun form will (e.g., She has a will of her own) strongly relate to individual volition, a central aspect of our embodied experiences as human beings. A fundamental human experience is desiring some state and intending to attain that state. For instance, a primary, universal experience is for an infant to feel hungry and desire to be fed; the infant’s hungry cry evidences her intention to attain the state of being fed. We would further argue that paying close attention to embodied meaning and its complex meaningfulness for humans is not the same as reliance on a general metaphor, such as time is motion. Indeed, our analysis relies on fine-grained embodied experience closely tied to internal volition, as labeled by will.

Our analysis of will is highly parallel to the analysis of the semantic extension of be going to from expressing a purely spatial meaning to an additional, independent future meaning put forward by Evans (Reference Evans2014) and Evans and Green (Reference Evans and Green2006). Evans argues that all lexical items are made up of a complex set of parameters, or atoms of meaning, that comprise their lexical sense. Resonating with Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) and Traugott and Dasher (Reference Traugott and Dasher2002), he argues that many of these parameters are inherent in the initial action labeled by the lexical unit. In the case of be going to, Evans presents bridging sentences (C. Johnson, Reference Johnson, Hiraga, Sinha and Wilcox1999), such as I’m going to eat, that can be interpreted as both expressing motion (i.e., the speaker has to move from their present position in space to another position in space in order to be appropriately placed to eat), as well as expressing the not-yet-realized intention to eat.

Evans further argues that the sense of prediction arises from intention. Finally, a prediction sense extends to a future sense. Various shades of meaning can be foregrounded by the contexts in which the word occurs. If a meaning is foregrounded frequently, it can become an independent sense associated with the lexical form (pragmatic strengthening). Thus, the meanings associated with be going to are argued to have evolved from spatial to intention to prediction to an independent expression of future that, in certain contexts, is no longer used to refer to movement through space. Importantly, Evans states that the invited inferences reside in the cognitive model or ‘semantic parameter that arises from motion scenarios’, thus moving the level of abstraction from the lexical level to a higher order schema. We believe this reasoning is entirely in the appropriate direction.

A key difference between our position and that of Evans (Reference Evans2015a), and indeed of Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) and Traugott and Dasher (Reference Traugott and Dasher2002), is that we argue that the declaration by the speaker of engaging in any process or action that has not yet reached an endpoint, entails not only an announcement of an intention, but also a prediction of fulfilling that intention. So, a sentence such as I’m leaving for the store invites the inferences that (a) the speaker intends to leave for the store; (b) the speaker expects (predicts) she will accomplish the act of leaving; and (c) the speaker intends and expects (predicts) she will arrive at the store. A similar set of inferences are present in a statement such as I’m giving a lecture in a few minutes. Indeed, these basic inferences are present in all scenes described by verbs of intentional action in imperfect aspect; they are grounded in everyday embodied experience. Thus, our notion of futurity would seem to come from multiple types of action (not just intentional motion from point A to point B); rather, a sense of future essentially stems from intending to do something, beginning the action, and understanding that, if completed, the action will be completed sometime after it was initiated. This sequence is highlighted when we communicate our intention to perform an action and thus declare our prediction that the intention will be realized.

In the next section, we argue that embodied experience is key to understanding the range of meanings associated with be going to and its semantic contrasts with will. Specifically, we show that the historical origin of go lies in Sanskrit and Old Germanic ga meaning ‘walk’. Moreover, the embodied experiences involved in the human walk cycle, as originally labeled by ga, rather straightforwardly account for the multiple meanings associated with current uses of be going to. Finally, we argue that this analysis is consonant with the LCCM model of semantics postulated by Evans (e.g., 2009, 2014, 2015b), in which he argues for semantic parameters that reflect embodied experience.

4.1. on the origin of go

In order to explicate the meanings associated with the construction be going to, we start by examining the etymological history of the word go, which is the lexical core component of the construction. Although be going to has clearly undergone a high degree of grammaticalization, and thus much of its original lexical meaning has been bleached away, recent studies show that many lexical units that have undergone grammaticalization continue to retain essential elements of the original lexical item (e.g., Evans, Reference Evans, Evans and Chilton2010; Mahpeykar & Tyler, Reference Mahpeykar and Tyler2015; Nikolaev, Reference Nikolaev2013; Sweetser, Reference Sweetser1990; Tyler & Evans, Reference Tyler and Evans2003), and be going to is no exception.

Sweetser (Reference Sweetser1990) argues persuasively that most modals, as well as many other words, retain key traces of their original meanings. Polysemy researchers argue that a given lexical item undergoing change will often “exhibit the co-existing layers of its past” (Tyler & Evans, Reference Tyler and Evans2003; see also Evans, Reference Evans, Evans and Chilton2010, Reference Evans2014). A look into the history of the verb go reveals that current meanings of be going to, which are often seen as difficult to explain as systematically related, can be accounted for when one takes its etymology into account.

Pérez (Reference Pérez1990) argues that the origin for go is found in Sanskrit jangha ‘the lower part of the leg’, as well as gan, which is analyzed as descending from a Proto-Germanic stem ga, meaning ‘walk’. Etymological dictionaries relate go to cognates in Old German (e.g., Hoad, Reference Hoad1986) and even present-day Dutch and Swedish, which mean ‘walk’.

4.2. on the constituents of be going to

Another component that is crucial to our understanding of be going to is consideration of all the constituents that make up the construction, e.g., the lexical verb go ‘walk’, the progressive marker -ing and the preposition to. Our analysis resonates with Bybee et al.’s (Reference Bybee, Perkins and Pagliuca1994) cross-linguistic discovery that general motion verbs tend to develop a future meaning only in the context of imperfective aspect and some kind of allative marker. Our analysis differs from Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) and Traugott and Dasher (Reference Traugott and Dasher2002) in two important ways. Traugott and Dasher summarize Bybee et al.’s explanation of why this type of construction develops a future meaning in the following manner: “motion towards something takes time (i.e. is imperfective) and one will arrive only at some time later than the motion starts” (Traugott & Dasher, Reference Traugott and Dasher2002, p. 83). Our sense is that this argument does not move us very far from Lakoff and Johnson’s (Reference Lakoff and Johnson1980) explanation of the metaphor time is motion; in both cases, this argument for humans’ sense of future rests on the co-occurring experiences of intentional forward motion through space and simultaneously realizing that time elapses from the point the motion begins until it ends. However, if we accept Langacker’s (Reference Langacker1987) basic definition of a verb, i.e., a verb indicates a process that exists through time, all verbs have the semantic quality of a process that finishes some time later than it began. The information that the process/action is taking place in time is part of what Evans (e.g., 2014) would call part of the general conceptual model of ‘verb’. Thus, all verbs would seem to have the potential to invite the inference of temporality; any verb used to express intentionality has the potential to invite an inference of futurity.

This raises the question of why, cross-linguistically and in English in particular, don’t just any intentional motion verbs develop a future sense? Notice that embedded in Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) and Traugott and Dasher (Reference Traugott and Dasher2002) is the notion that the motion verbs that give rise to a future marker are verbs of forward motion. Indeed in English, a number of verbs of forward motion are used to indicate a not-yet-realized end-state or goal, and hence an implied future:

1. (13) a. We are rapidly moving towards a conclusion.

   b. We are crawling towards the end of the semester.

But these verbs of forward motion have not developed as independent future markers.

Our hypothesis is that there is something fundamentally more salient about the general, default human action of moving one’s entire body from point A to point B that makes words indicating ‘coming’ and ‘going’ more likely to develop a future meaning. For most human beings, the major act of intentional motion is moving the whole body from one location to another, specifically walking. After all, standing erect and circumlocating via our hind legs is one of the anatomical design features that make humans unique. Moreover, walking is a highly salient achievement most infants accomplish during their first year of life, absorbing many hours of concentrated experimentation and practice; it’s not surprising that this general motion may be more salient than other intentional motion, such as reaching out one’s hand and grasping an object as a reflex movement.

The information the imperfective adds is that the action is not yet completed. It is that subjective focus, in which the endpoint of the process is not part of the scene being conveyed, that signals that the endpoint has not yet been achieved, and hence strongly implicates the potential for some future completion. In general, humans interpret each other as rational creatures who undertake an action in order to reach some sort of goal or endpoint. However, the allative marker explicitly marks and therefore highlights an intended goal. The imperfective and the allative marker combine to paint a scene focusing on the not-yet-realized status of the intended action. In sum, the progressive directly signals that the subject is in the midst of the activity, and thus that the endpoint is not yet actualized; the prepositional marker to signals directionality, pointing to the goal (Tyler & Evans, Reference Tyler and Evans2003), and so also contributes to the not-yet-actualized semantics of the construction (the fact that to is used as the infinitive marker in English underlines its contribution of ‘not yet actualized’ generally in the English language and, in particular, in this construction).

4.3. be going to , embodied experience, and the walk cycle

Recent work on the inter-relatedness of human language, cognition, and the human bodily experience has shown that our everyday interactions with the world shape our cognition as well as our language (e.g., Bergen, Reference Bergen2012). An abundance of studies have shown that embodied experience plays an important role in explaining why certain meanings are associated with certain words (e.g., Bergen, Reference Bergen2012; Evans, Reference Evans2014; Gibbs, Reference Gibbs2003). If we ignore embodied experience and etymological origins, the picture may not be clear as to how many of be going to’s meanings could have evolved from the general, present-day meaning of go as a mannerless motion verb. In the absence of this consideration, Langacker (Reference Langacker1987, Reference Langacker1991) and Brisard (Reference Brisard2001) have paired go’s meaning of mannerless motion to the vague, general metaphor time is motion and rejected the metaphor’s ability to explain the multiple meanings associated with be going to, especially for its inadequacy to explain how meanings like prior intention or inevitability came to be associated with the construction (Brisard, Reference Brisard2001).

Although Brisard (Reference Brisard2001) correctly identifies the notion of prior intention as a central component of the current meaning of be going to, he argues that this meaning is not part of the original meaning of the construction. He further argues that a goal of his analysis is to determine how prior intentionality came to be associated with be going to, which he does by positing +P, i.e., speaker’s access to confirmatory evidence in the present. In this particular case, the speaker has confirmatory evidence that the action will occur based on having already planned to undertake the action. However, Brisard does not explain the process by which be going to, which did not originally have the +P feature, came to have one; nor the process by which will, which very well could have had a +P designation, lost it.

When one considers the original meaning of walk, the picture becomes clearer. The action originally indicated by go, with its historical roots in Sanskrit and Old German and its relation to current Danish ga ‘walk’, involves a good deal of intentionality. Indeed, intentionality is a central characteristic of the human act of walking, as we discussed above, simply not found in many verbs of movement, such as twitch, jerk, fall, tumble, etc.

We contend that prior intention associated with be going to arises as a straightforward invited inference stemming from the act of walking; intentionality is and has always been a central component of the human act of standing erect and moving forward, i.e., in the conceptualization and performance of walking. In Evans’ (Reference Evans2015b) terms, prior intention is part of the cognitive model associated with walking. For the developing infant, learning to stand is a long, intentional, effortful process involving learning how to resist gravity and find balance. Certainly, the development between the self-locomotion of crawling and the self-locomotion of walking is highly intentional. Although at certain moments the young child may appear to be unaware that she has actually taken steps, the many attempts that allowed the child to reach this stage were self-aware and intention-ful. When the young toddler wants an object that is at a distance, the act of walking in order to reach that object is clearly intentional. Once the individual learns to walk, the muscle memory becomes deeply entrenched and in many ways is automatic. But unlike the automatic alpha wave cycles or the pumping of one’s heart, walking is an ‘automatic’ action under the control of the individual. One decides whether to remain sitting or move to a standing position; one decides to move from a static standing position to the motion event of walking.

Further, we argue that the other meanings associated with be going to are motivated by human embodied experience of walking and its kinesthetic dynamics. The action of walking and the spatio-physical interactions and affordances created by walking create experiential meanings that we identify with this action. Walking takes volition and prior planning, which involves lifting a leg with the support of core muscles and arms, as well as other gross motor muscles participating in balancing the body and keeping an upright body posture. Once the walk cycle is initiated, an inevitable sequence of events occurs. One leg is lifted and the body makes a forward motion, with the opposite arm moving forward in synchronization; it is inevitable that the lifted leg will come down in momentum (it is inarguable that a human cannot stand on one leg indefinitely) and the arm that was swinging forward will swing back like a pendulum caught in motion. As the first leg goes down and touches the ground, with the body caught in the inertia of motion, the second leg lifts and the opposite arm moves forward. Once walking is established as a routine activity of the young child, a sense of inevitability becomes closely tied with this physical progression and momentum of the force dynamics involved in walking. ^{Footnote 4}

In terms of imminence and assumption, an infant learning to walk will learn, as she takes more steps in the world, that the distance that separates her from objects she wishes to obtain will shrink as she takes forward steps, bringing her closer to a parent she is trying to reach or a desired toy she wants. Entities in her surroundings become closer, physical cues in the environment become more apparent, and her sense of accomplishing her goal becomes imminent. As the infant takes more steps, she will develop concepts about affordances and the amount of effort needed to reach her goals. As she becomes an expert walker, she will become better at observing possible barriers to her forward progress, better at figuring out how to overcome these potential barriers, and thus better at making accurate estimates about how reachable the objects in her environment are.

4.4. on the english modals and embodied experience

The role of embodied experience and force dynamics has been used to explain many linguistic phenomena, including the meanings of the English modals. For example, Sweetser (Reference Sweetser1990) argues that while on the surface the epistemic and root uses of the English modals seem to bear no similarities in meanings, the link between root and epistemic uses of modals can be systematically explained by means of defining the modals in terms of appropriate force dynamics (forces of propulsion and barriers to forward movement) and extension through conceptual metaphor. In other words, Sweetser suggests that the link between the two types of modal use lies in metaphorically extending our understanding of exterior, social and physical forces and barriers – in the case of root modality – to internal, mental forces of logical reasoning – in the case of epistemic modality. Sweetser does not directly talk about these forces in terms of embodied experience, but notions such as social and physical forces exerted upon an actor and internal forces propelling an actor strongly imply embodied meaning; these explanations are, after all, grounded in affordances and interactions humans have with the real world. Tyler’s (Reference Tyler, Robinson and Ellis2008) analysis more overtly adds embodied experience to Sweetser’s analysis by representing the force dynamics associated with modals in term of human actors and human levels of certainty about as-yet-unrealized events and states, as well as distinguishing the modals in terms of the particular types of forces propelling the actors and barriers to the actor’s forward movement. R. Lakoff (Reference Lakoff1972) refers to will as the strongest modal that shows the strongest commitment to certainty of future action. This analysis involves the actor’s internal commitment in the present to an unquestioned future event or state. Sweetser (1990, p. 55) observes that in many utterances involving will-futurate “the speaker undertakes to see to it or to command that the action be done” (an external authority, such as the law, can exert this responsibility or commanding power). Our position is that the central future reading of will involves that speaker’s declaration of her strong internal determination to complete an action or intention. As discussed above, an inherent invited inference to such strong declaration is the speaker’s strong prediction of the successful completion of the intended action. In physical-force dynamic terms, the speaker is declaring an absence of any foreseeable barriers to completing the action or intention; the path to the future is clear, level, and unobstructed. Sentences (14a–b) illustrate the root/future and epistemic meanings respectively:

1. (14) a. Karen says she will finish the project this morning.

   b. [Some unseen person knocks at the door] That will be Karen.

Through repeated highlighting of the inherent, invited inference of strong prediction, an understanding of the root force dynamic progression (strong determined forward motion with no barriers to realizing the desired action or state and a declaration of strong intention with the concomitant strong prediction of success) is extended to an epistemic version of will. In the sentence That will be Karen, the speaker identifies with absolute certainty that the unverified hypothesis that Karen is the unseen person knocking at the door will be confirmed; there is strong evidence to support the speaker’s hypothesis and no foreseeable alternatives to stand in the way of drawing that conclusion.

Tyler (Reference Tyler, Robinson and Ellis2008) drew on Sweetser’s (Reference Sweetser1990) explanations of the English modals involving force dynamics and – by implication – embodied experience, to posit visual representations of the semantics of the modals. Where will is concerned (see Figure 1), Tyler represents the meaning as a figure, the actor or mover, who is moving forward; the lines in the figure’s head indicate internal force, volition, or desire. The figure’s double arms are extended forward to represent strong forward momentum and determination. There are no barriers in the figure’s path.

Fig. 1. An embodied, force dynamic representation of will reproduced from Tyler (Reference Tyler, Robinson and Ellis2008).

Following Tyler’s (Reference Tyler, Robinson and Ellis2008) model, we propose a systematic representation that captures the meanings of be going to within this general framework for the English modals and periphrastic modals. For the first meaning, prior intention, we represent a ‘ghostly’ actor in the position that precedes the moving figure; the ‘ghostly’ figure represents the intention or planning that precedes the movement, distinguishing between will’s meaning of strong, internally generated determination and be going to’s prior or planned intention (see Figure 2). As explained earlier, walking requires an intentional choice or plan that precedes the physical movement. The category of prior intention corresponds to the source of motion or energy behind it, i.e., the prior decision or plan to walk. Figure 2 illustrates prior intention as the figure in broken lines that precedes the figure in bold lines, where the latter represents the present moment of speaking. The future (represented by the shadowed area) is projected from this prior intention: although not yet realized, it is strongly presumed, and it provides a central sense that is relevant in the description of all usage types of be going to.

Fig. 2. Prior intention.

As explained earlier, prior intention is an inherent part of the human embodied experience of walking. It is this inherent meaning that makes sentences such as the one in (16) sound ungrammatical or unacceptable (after Catasso, Reference Catasso2012).

1. (15) a. I can’t stand cigarette smoke.

   b. Oh, I’ll stop smoking this cigarette.

2. (16) a. I can’t stand cigarette smoke.

3. b. *Oh, I am going to stop smoking this cigarette.

Such usage, where intention does not originate prior to the time of speaking, is acceptable with will but not be going to.

The second meaning is that of imminence, where future is presented as near and immediate. If we were to think of an actor in the midst of walking towards a goal, the actor will have a clearer perception of a target than when he started. As one physically approaches the goal, objects, states of affairs, or events that were distal are seen to be nearer, thus providing physical signs of the successful end of the action approaching. As the actor navigates through the physical world, more immediate evidence in the environment becomes visible or more clearly apprehended. Figure 3 illustrates the meaning of imminence (based on observation of cues), where the observed, tangible, and physical clues are represented by the eye in the actor’s head.

Fig. 3. Imminence.

The third meaning attributed to be going to is the meaning of assumption. This meaning is similar to that of imminence in that the two categories draw on clues and signs to the future state of affairs. Brisard (Reference Brisard2001) argues that the primary difference between the two lies in the type of clues available. In the case of imminence, the cues are, for the most part, immediately present in the environment. In contrast, in the case of assumption, the cues stem from the conceptual knowledge in the mind of the speaker. In both cases, evidence is a resource on which the speaker draws conclusions about the future and assumes the interlocutor is likely to reach the same conceptualization. If one has walked a route numerous times, a set of memories or conceptualizations about what to expect are built up. For instance, if one walks a route that often has lots of traffic, one learns to take care when crossing the road. Even on a particular day when the traffic is light and one does not have the immediate auditory and visual cues of oncoming traffic, one continues to be alert and assume, based on past experience and conceptual knowledge, that caution is necessary. Not-yet-encountered traffic is likely to appear. Thus, a store of conceptual knowledge related to walking various paths is experientially correlated with the walk cycle. Figure 4 illustrates the notion of conceptual clues that project the future state. The source of this projection is the mind of the conceptualizer.

Fig. 4. Assumption.

The last meaning associated with be going to is that of inevitability. It is represented by the progression of the leg movement in the walking action. As discussed above, once the walk cycle is started, certain movements, like the swinging of the arm and the placement of the raised foot back on the ground, are inevitable. With this meaning, the future is represented as proceeding with a very high degree of certainty and force. Figure 5 illustrates this meaning.

Fig. 5. Inevitability.

In this series of overlapping figures, the first figure is in solid lines and is depicted with one leg raised; this is the first physical stage of the walk cycle. The second figure is in dashed but bolded lines, representing the inevitable, very near future motion that immediately follows the initial action. Note that the lifted leg is descending relative to the height of the leg in the initial position. The third figure is in dashed, unbolded lines, indicating a position in a slightly more removed future. Here the leg is in contact with the ground; the first half of the walk cycle is completed. The overlapping of the three figures indicates that the action is continuous.

Under our analysis, then, the overall meaning of futurity associated with be going to comes from the information supplied by the full construction that the actor is in the midst of an action (signaled by the imperfective/progressive) of forward motion involving the entire body (signaled by the lexical unit go ‘walk’) towards a goal (signaled by to). In many cases, the not-yet-realized goal will be a not-yet-realized action, event, or state. When part of a usage-event, this compositionally formed meaning strongly invites the inference that the goal (or action, event, or state) is highly likely to be reached at some later time, i.e., the future. Additionally, the particular subjective stances towards the not-yet-realized event are a set of invited inferences inherent in the embodied experience of walking. In other words, the extended meanings of prior intention, inevitability, imminence, and assumption that distinguish be going to from will all arise from invited inferences inherent in a cognitive model linked to the original lexical item, [ga].

The spirit of this analysis is coherent with Evans (e.g., 2009, 2010, 2014, 2015b) in that it recognizes invited inferences as (nearly) synonymous with the notion of each lexical unit being semantically complex and comprised of multiple cognitive parameters, or semantic atoms. Through pragmatic strengthening, certain of these component semantic elements can be highlighted repeatedly until they are established as independent senses associated with the lexical item. Our analysis also emphasizes embodied experience associated with usage-events as part of the process of semantic extension. This is the missing link in the semantic extension arguments put forward by Bybee et al. (Reference Bybee, Perkins and Pagliuca1994) and Traugott and Dasher (Reference Traugott and Dasher2002). While the declaration of a speaker’s intention to do something involves an important invited inference that accompanies the usage event of uttering a sentence such as I’m going to the store, concomitant with the declared intention is the speaker’s prediction that the intention will be realized. The final step in the establishment of the future markers is that such predictions are regularly realized. Thus, a strong mental association is formed between the act of declaring an intention and concomitantly the actual realization of that intention. Futurity seems to be an invited inference inherent in all usage-events involving the speaker’s declaration of an intentional action. We hypothesize that will developed as the default marker of futurity from the very strong declaration of desire and volition inherent in its central meaning. This suggests that the widely accepted explanation that human understanding of future resides in our experiences of physical/spatial forward motion towards a goal simultaneously taking place in time has been overstated (e.g., Bybee et al., Reference Bybee, Perkins and Pagliuca1994; Evans, Reference Evans2014; Lakoff & Johnson, Reference Lakoff and Johnson1980; Traugott & Dasher, Reference Traugott and Dasher2002). On the other hand, be going to obviously has to do with moving from point A to point B (with point B being highlighted as a goal) and did develop into an important marker of future in English. We have argued that the particular lexical item go developed as the central lexical item in this future construction because the verb originally labeled the particularly salient human action of walking. Thus, we would not deny that some of our human understanding of time, and in particular the future, stems from our embodied experience of moving through space, coupled with an understanding of time elapsing as we move from point A to point B. However, as Evans (Reference Evans2004) elucidated, our conceptualization of time is far more complex and nuanced than this.

Evans’ LCCM model addresses three kinds of polysemy, including what he terms ‘inter-lexical polysemy’. His groundbreaking (2010) chapter makes the first persuasive case for the need to recognize and account for the phenomenon of two (or more) lexical items developing broadly overlapping meanings that are nevertheless distinguished in subtle ways. Under Evans’ analysis, these subtle differences arise from the original parameters of the distinct lexical items. Evans (Reference Evans, Evans and Chilton2010) notes that both in and on have developed extended ‘state senses’, as in Chuck’s in love and The police are on high alert. However, corpus data shows that the two prepositions occur with a distinct set of collocates and the nuances of the states indicated by in are distinct from those indicated by on. He further argues that the semantic distinctions can be traced to the distinct parameters of the cognitive models associated with the original spatial/functional meanings of the two prepositions. We argue that will and be going to are also examples of inter-lexical polysemy. They both share a general meaning of future; however, close examination reveals that, in the majority of cases, the two forms have subtle differences in meaning. Moreover, these differences in meaning can be traced back to differences in the original meanings (or the original cognitive models) associated with the lexical items.