2. A preliminary assessment of the constructions under analysis

Let us first focus our attention on the transitive resultative patterns illustrated in (1)–(2) below. More specifically, these can be described as goal-oriented transitivity patterns that designate the outcome of a change of state (with or without a motion ingredient) undergone by the patient or affected object:

1. (1) Mabel wiped the table clean. (COCA, 2003)

2. (2) Men danced themselves into a frenzy. (GBAC, 2008)

As briefly mentioned in the ‘Introduction’, the notation employed here is the following. Examples such as They laughed themselves silly (COCA, 2000), He hammered the metal flat, etc. are labeled AP RCs, irrespective of whether they are exploited figuratively, as in the case of ‘laugh oneself silly’, or non-figuratively, as in ‘hammer the metal flat’. Examples like He broke the vase to pieces, They danced themselves into a frenzy, etc., where the change of state is invariably perceived as a metaphorical change of location, are termed PP resultatives. The reader may note that the PP in this variant of the RC is but a figurative use of the motion PP of canonical (caused) motion constructions (e.g., Jimmy pushed the boat into the water (COCA, 1999)). Both the AP and PP versions of the RC will collectively be referred to herein as ‘resultatives’, since both designate the endpoint of a change of state (cf. Goldberg, Reference Goldberg1995, p. 181, for a similar view). PP resultatives will in turn be distinguished from the CMC, which involves the actual or prospective change of location by an instigator of motion in such a way that an entity moves along a path (cf. Goldberg, Reference Goldberg1995, p. 152). As was the case with AP resultatives, the CMC may be figuratively exploited, as in They laughed the actor off the stage, which conveys self-instigated motion as a result of psychological impact being metaphorically understood in terms of physical impact, or they may refer to literal, externally caused changes of location (e.g., He kicked the ball into the net) (Ruiz de Mendoza & Luzondo, Reference Ruiz de Mendoza, Luzondo, Brdar, Raffaelli and Zic Fuchs2012; Ruiz de Mendoza & Mairal, Reference Ruiz de Mendoza, Mairal, Radden, Köpcke, Berg and Siemund2007, Reference Ruiz de Mendoza and Mairal2008).

Semantically, PP resultatives and the CMC differ in that whereas in the former the patient undergoes a change of state (expressed through figurative motion), in the latter the result element is that of a change of location. Despite slight formal and semantic divergences, all these instantiations form part of a broader family, i.e., the family of the resultative, first studied by Goldberg and Jackendoff (Reference Goldberg and Jackendoff2004) (cf. Boas, Reference Boas2005, for a critical revision).^{Footnote 7}

Terminological issues aside, even though Saurenbach (2008, p. 227) is “hesitant to give resultative and directional phrases a unified conceptual treatment”, we believe that from the point of view of the semantics of the RC, a unified approach in which the resultative element can take the form of both an AP and a PP is desirable. In addition, authors like Boas (2008, p. 128) assume that, especially in the case of RCs, it is difficult, if not almost “impossible, to predict a verb’s distribution of arguments based on a construction’s capability of contributing arguments to the semantics of a verb”. In this view, verbal predicates tend to exhibit idiosyncratic behavior when incorporated into the resultative syntactic frame, often being reluctant to subtle lexical or syntactic changes, as shown in (3):

1. (3) a. He drank himself {into a coma/sick/?ill/?comatose/?into sickness}.

   1. b. He hammered the metal {flat/into a flat sheet/?to flatness/*long/*tubular}.

However, rather than being an arbitrary phenomenon, at the end of this section we argue that the selection of an AP or a PP in RCs is motivated by general tendencies or principles.

The reader will concur on the fact that, depending on factors such as the nature of the affected object, the type of activity carried out by the subject referent, etc., entities can undergo various kinds of transformations, thus resulting in specific changes of state. We may thus expect language to reflect an array of (conceptually akin) syntactic patterns capturing such diversity (cf. Radden & Dirven, Reference Radden and Dirven2007, p. 391). By way of illustration, consider the following sentences:

1. (4) a. The lake froze solid.

   1. b. Richard drank himself silly. (COCA, 2010)

   2. c. Richard ate himself sick. (COCA, 2000)

   3. d. Erin talked herself hoarse. (COCA, 2012)

   4. e. The blacksmith hammered the metal flat.

   5. f. The new owners painted the front door red. (COCA, 1996)

In these sentences, an (explicit or implicit) agent acts on a patient that, as a result of the activity denoted by the verb, changes its original state by acquiring a new property: in (4a) all or at least a significant part of the water in the lake is now solid owing to the effect of low temperatures; in (4b) Richard was acting in a silly way as a result of the influence of excessive alcoholic consumption; in (4c) Richard’s excessive eating caused him to be sick; in (4d) Erin talked so much that she became hoarse; in (4e) the metal became flat as a result of the blacksmith hammering it; in (4f) the novel state of the door being red results from the action of painting it. Although superficially similar, there are yet differences among the various realizations of the AP resultative presented in (4). One such difference has to do with the temporal duration of the outcome events. Some, as in the case of (4a)–(4d), are only transitory changes of state, while others, as in (4e)–(4f), are permanent. In other words, eventually, the affected entities in the first four examples will naturally go back to their initial state. By contrast, in the default scenario, the resulting features of being respectively flat and red in (4e)–(4f) will always prevail unless the objects are again manipulated and therefore caused to change. Whether the resulting state is permanent or temporary, we want to draw the reader’s attention to the fact that when patients are partially affected (i.e., only a new property or a set of new properties, such as a different shape, color, consistency, is acquired), an AP tends to be employed. Some additional examples illustrating this claim are: The joggers ran their Nikes threadbare (given in Boas, Reference Boas2003, p. 47), She wiped the floor dry with a floor rag (COCA, 1996), I’m going to slap you silly (COCA, 2008), He drank himself blind (GBAC, 2003), Jeff falls to the ground and hits his head knocking himself unconscious (COCA, 2007), He pulled his finger out of the box and licked it clean (COCA, 2005), His father scraped the boot clean (GBAC, 2011). As with the examples in (4) above, in none of these realizations does the essence of the patient change substantially: the floor is now dry, although it is exactly the same entity, a person’s finger does not suffer any conspicuous change just because it is licked clean, etc.

But AP resultatives may also vary in relation to the period of time spent in accomplishing a (usually conventionally expected) result (cf. Boas, Reference Boas2003), which often hinges on the nature and kind of affected objects involved in conjunction with the type of activity in which they are engaged. The following examples bear this point out: He drank his glass empty (COCA, 1999) < He hammered the metal flat < They painted the house green.^{Footnote 8} In default scenarios, whereas one can drink the liquid contained in a glass in a few seconds, it usually takes longer to make a piece of metal flat by hammering on it, and it even takes longer to paint a whole house. Note, additionally, that the verbal and the constructional subevents are not necessarily coextensive (see Goldberg & Jackendoff, Reference Goldberg and Jackendoff2004, pp. 545–546, for a discussion). While in ‘drinking a glass empty’ we tend to think of the state of being empty as the final stage of the drinking event, in ‘hammering a metal flat’ and ‘painting the house green’ the states of being flat and green can more easily be seen as temporally co-extensive with the verbal action than becoming empty. The reason for this lies in the fact that a glass is not empty until its contents have been fully finished, while metal can gradually achieve greater and greater degrees of flatness and one can say that a house has been painted green even if still not fully coated with paint. The same rationale applies to (4a), since, according to Iwata (2006, p. 458) the events of freezing and becoming solid “are co-extensive and unfold at the same time”.

Yet more importantly for our purposes here is the fact that divergences also occur between low-level pairs employing an AP and those adding motion to the state of affairs:

1. (5) a. He bought a bottle of whiskey (…) fully intending to drink himself asleep. (GBAC, 2004)

   1. b. He drank himself to sleep. (COCA, 2005)

2. (6) a. Jensen broke the cage open. (COCA, 2000)

   1. b. The glass or vessel being empty, the bridegroom throws it on the ground and breaks it to pieces. (GBAC, 2003)

3. (7) a. The blacksmith hammered the metal flat.

   1. b. We hammered hot iron into knives. (COCA, 1991)

Before examining these examples, there are two relevant issues that need to be pointed out. First, the crucial feature setting apart the sentences employing an AP and those resultatives including a motion preposition is that only the latter build on A CHANGE OF STATE IS A CHANGE OF LOCATION (cf. Lakoff, Reference Lakoff1987, Reference Lakoff and Ortony1993; Lakoff & Johnson, Reference Lakoff and Johnson1999), a high-level metaphor which we discuss in more detail in the following sections. However, this does not mean that AP resultatives, like the CMC, are not motivated by other high-level metaphors and metonymies, as shown in Section 4. Second, it is interesting to note that the choice of a PP over an AP is, in many cases, dictated by the fact that English has no adjective available that can capture the intended meaning, and thus speakers resort to the use of the more productive category of PP resultatives.^{Footnote 9} With this in mind, let us now discuss each of the examples above.

Whereas (5a) focuses on the result, the use of a motion ingredient in (5b) gives greater prominence to the process that leads to the result. An additional meaning effect that arises from this difference in prominence is the idea that the event in (5b) appears to be longer than that of (5a). This meaning effect is grounded in the exploitation in (5b), but not in (5a), of the metaphor whereby we see changes of state as if they were changes of location. Generally, a change of location involves motion along a path from a source to a destination. Although, admittedly, the distance between source and destination can be covered almost instantly, our common everyday experience, on which the metaphor is based, tells us that covering distances (especially at human and animal ground speeds) takes time. Compare, by contrast, the examples given in (6), both of which codify instantaneous results. However, there is a clear difference between them. When one breaks something open (e.g., (6a)), the affected object does not turn into a fundamentally different entity. Yet the same is definitely not true of (6b), in which the object is fragmented to a point in which it can no longer be considered the same entity. In turn, the use of a PP in (7b) is comparable to that of (6b), in the sense that in both instances there is a radical, holistic transformation, which results in the entities ceasing to be what they were before the action was performed upon them. Nonetheless, this is not necessarily true for all resultatives involving figurative motion, as attested by the sentence Steven had worked himself to exhaustion to earn his master’s degree (GBAC, 2005). That is, although in this last example, reaching the state of exhaustion is also codified through the metaphor A CHANGE OF STATE IS A CHANGE OF LOCATION, the subject does not undergo a severe transformation of the type expressed by other realizations of the PP RC such as those in (6b) and (7b) above. Finally, the change in (7b) is more salient than the one depicted in (7a), given that, under normal circumstances, turning a piece of metal into a knife requires a more elaborated process than simply banging it flat.

From this analysis, we conclude that, when metaphorical motion is used to express a change of state, the change is perceived as more conspicuous than when directly coded into an AP. This is only true, of course, provided that one such AP is available in English. If no such AP is available, the only remaining option is a PP, as in He begins hammering it into the shape of a horseshoe (GBAC, 2002). There are, therefore, three interpretation possibilities: (i) the change may result from an event that is perceived as being prolonged in time (e.g., He had worked himself into the position of acting governor; GBAC, 2007); (ii) it may entail a more elaborate process (e.g., Spaniards obtained (…) pieces of gold worked into the shape of fishes or birds; GBAC, 2004); or (iii) it can indicate a complete transformation (as evidenced by such sentences as The witch turned the prince into a frog, The butter melted to liquid, He smashed the vase into tiny pieces, etc.).

The non-exhaustive range of constructional realizations presented so far evidences the need to postulate a continuum along which RCs are categorized on the basis of an increasing change parameter with various degrees of affectedness: non-figurative changes based on the sole acquisition of a property or set of related properties (e.g., The river froze solid, Tom kicked Paul black and blue), changes of state based on figurative motion in which the essence of the patient is not altered (e.g., He drank himself into a stupor (COCA, 2009)), or changes of state involving figurative motion that result in a complete transformation (e.g., Should I see his body before it is burnt to ashes? (COCA, 1990)).

Before closing this line of argumentation, recall that at the beginning of the present section we contended that the selection of a resultative phrase in RCs, far from being an inconsistent phenomenon, is a principled one. More specifically, we contend that it gravitates around the following three principles (cf. Ruiz de Mendoza & Luzondo, Reference Ruiz de Mendoza, Luzondo, Brdar, Raffaelli and Zic Fuchs2012):^{Footnote 10}

1. a. If the affected entity ‘A’ is perceived or construed as experiencing a holistic, radical change so that ‘A’ can no longer be considered the same entity, a PP is required: The ice melted to liquid, She painted the brush to pieces. ^{Footnote 11} We shall symbolize this situation as A>B, where A and B denote different entities.

2. b. However, if the affected entity ‘A’ acquires a new property/quality but ultimately retains its essence, an AP is employed: She drank herself blind, They painted the house green. We shall label this situation A>A´, where the prime mark denotes a new property.

3. c. Between the two ends represented by A>B and A>A´, a third category can be posited, which we shall call the ‘motion’ A>A´ change of state of pattern. This category encompasses figuratively exploited resultatives which are formally PP resultatives but semantically do not codify the same kind of change as that of pure A>B realizations (e.g., People may eat themselves into a heart attack (GBAC, 2012)).

To close this section, let us consider some examples that fall within the scope of the motion A>A´ pattern:

1. (8) a. The vegan parents starved the kid to death. (COCA, 2007)

   1. b. In 1968, lightning knocks Wanda into a coma. (COCA, 2006)

   2. c. (A)nd other barnyard subjects that bored me into a stupor. (COCA, 1994)

   3. d. She drank herself into a state of paralysis. (COCA, 2006).

In none of the examples in (8) is the essence or the core features of the patients altered. As opposed to realizations like New York museum burned to ashes in 1968 (COCA, 2006), It [the bowl] falls to the floor and smashes to bits (COCA, 1990), the affected entities in (8) are the same whether alive or dead, in or out of a coma, a stupor, etc. In consequence, the A>B pattern (cf. The witch turned him from a prince into a frog) does not apply to (8a)–(8d) even if the metaphor A CHANGE OF STATE IS A CHANGE OF LOCATION is also operational. Yet the examples in (8), despite being subsumed under an A>A´ pattern, are not fully comparable to those utilizing an adjectival resultative phrase (cf. He hammered the metal flat: A>A´), since, as discussed in (5)–(7) above, when motion is added to the state of affairs, the focus falls upon the process leading to a result, thus bringing about different meaning effects.

As previously pointed out, however, there may be a simpler reason for the choice of a PP instead of an AP. In line with Boas (2005, p. 451), prototypical adjectival resultatives are only partially productive. For this reason, if there is no adjective available to codify the intended meaning (e.g., *She bored me stupored) the change of state is necessarily realized through a directional. In some other cases, both an AP and a PP may be used (e.g., dead / to death). However, in (8a), for example, employing figurative motion is consistent with the need to express an event that develops in time until the result (i.e., causing somebody to die from starvation) finally obtains. The string “starv* * (*) dead” yields no results in the COCA. Seemingly, when shoot collocates with an AP (e.g., Molly shot him dead (COCA, 2010)), the result is more instantaneous than when a PP is selected (cf. Goldberg & Jackendoff, Reference Goldberg and Jackendoff2004, p. 561).

In view of these uses, we now address the relation holding between the RC and the CMC, which Goldberg (Reference Goldberg1995) argues to be a matter of metaphorical extension.

3. Metaphorical extension revisited

The connection between the prototypical AP resultative (e.g., He hammered the metal flat) and the literal CMC (e.g., Pat threw the book off the table) has been the object of some debate. Goldberg (1991, 1995, p. 81), for example, claims that the former construction is a metaphorical extension of the latter since the resultative element of AP resultatives can be interpreted as a metaphorical type of goal. In Goldberg and Jackendoff (2004, p. 542), it is further contended that, because AP resultatives are based on PP resultatives, “we should not find languages which have only AP resultatives and not PP resultatives”. Peña (2009, pp. 748, 757), who partially takes sides with Goldberg (Reference Goldberg1995), introduces some degree of refinement into the issue by calling for a cognitive continuum between the two. At the ends of such a cline we have canonical examples like He hammered the metal flat for the resultative and Frank pushed it into the box for the CMC. Somewhere in the middle of this continuum we find less central cases (e.g., The kid broke the vase to pieces, They laughed him out of the room, Someone persuaded her into love, etc.), which resemble more or less faithfully the above-mentioned archetypes. Boas (2003, pp. 6, 94–99), who groups caused-motion examples like Frank sneezed the napkin off the table or Dobson laughed you off the street under the label of RCs, argues against the analysis of the RC being a metaphorical extension of the CMC. In turn, Hampe (2010, p. 187) claims that subsuming both object-related adverbials and object-complements into one single category downplays the differences between the resultative phrases. She further gives support to the distinction between the RC and the CMC through a collexeme analysis (cf. Stefanowitsch & Gries, Reference Stefanowitsch and Gries2003) and argues, on the grounds of a quantitative study exploiting the Child Language Data Exchange System (CHILDES), that the Goldbergian claim that there is a metaphorical inheritance link between the CMC and the RC is implausible from a developmental perspective. When justifying this point, Hampe (2010, p. 206) observes that complex transitive “expressions with resultative verbs […] occur before metaphorical instances of the CMC and before instances of caused-motion verbs in the RC”.^{Footnote 12} This means that people acquiring their mother tongue are more likely to construct an AP resultative like He hammered the metal flat before they learn to use PPs figuratively to express result (He hammered the metal into different shapes); and last of all, they learn to construct sentences like She drove me into despair, which combine the use of a non-resultative predicate with the expression of result through figurative caused motion. Hampe’s study definitely contradicts Goldberg’s hypothesis on empirical grounds: the RC is not patterned after the CMC, but the other way around. We will now add linguistic evidence in the direction of Hampe’s empirical claims.

Goldberg (1995, pp. 81–89) contends that the prototypical AP resultative involves a metaphorical interpretation of the resultative component as a figurative type of goal, through the metaphor A CHANGE OF STATE IS A CHANGE OF LOCATION. Following Goldberg’s rationale, because the adjectival resultative phrase codes a metaphorical change of location, a wide variety of co-occurrence restrictions can be accounted for. A case in point is the fact that resultatives cannot occur with directional phrases regardless of sequence:

1. (9) *Sam kicked Bill {black and blue out of the room / out of the room black and blue}.

For Goldberg, (9) is ungrammatical in the light of the Unique Path Constraint. According to this constraint, a single clause cannot specify more than one distinct path for the same constructional argument, that is, an argument “cannot be predicated to move to two different locations at the same time” and “motion must trace a path within a single landscape” (Goldberg, Reference Goldberg1995, p. 82). In her view, black and blue codes a metaphorical change of location that specifies a path distinct from that of the non-figurative change in space (i.e., out of the room). In other words, admitting the existence of two distinct paths is tantamount to saying that the object referent is moving to two distinct locations at the same time, which is impossible.^{Footnote 13} In Goldberg’s (1995, p. 83) own words, resultatives are “restricted from occurring with a directional because the directional, coding a change of physical location, will code a path distinct from that coded by the change-of-state resultative”. From our point of view, it is hard to assume that black and blue presupposes any kind of figurative motion. This would mean that this resultative expression is a metaphoric source involving motion, which is obviously not the case.^{Footnote 14} By contrast, figurative motion is operational in resultative examples taking the form of the motion A>A´ pattern (e.g., I punch you and kick you into submission (GBAC, 2010)), or the A>B pattern (e.g., He kicked it (the guitar) to pieces (GBAC, 2007)). It should be borne in mind that there is a tendency to use APs when patients are only partially affected, usually acquiring a new property or a set of new properties: They painted the house white (A>A´), His plate looked like a dog licked it clean (GBAC, 2005: A>A´). Following this rationale, in Goldberg’s example, i.e., Sam kicked Bill black and blue, Bill simply becomes bruised as a result of the action denoted by the verb, thus naturally falling within the A>A´ pattern, which depicts changes in the absence of any motion ingredient.

As far as the impossibility for the co-existence between black and blue and out of the room in the same clause, this does not have to do with the fact that the Y constructional argument – following the Goldbergian notation – is both metaphorically and literally construed as moving to two distinct locations, as Goldberg advocates. Rather, its ungrammaticality simply arises from the fact that once the Z slot has been filled with a change of state (i.e., Y becomes bruised), further changes of any other kind are blocked in order to avoid redundancy. This is so because both black and blue and out of the room convey different results. For example, whereas in He kicked Paul black and blue, the patient experiences a change of state without motion being involved (i.e., ‘Paul becomes bruised’), in He kicked him out of the house, the patient literally ends up in a different location. Nevertheless, as will be shown at the end of this section, while the conceptual proximity between changing states and changing locations impedes having black and blue plus out of the room in the same sentence as resultative elements of the same verbal predicate, it is precisely such conceptual proximity that allows speakers to construe the destination of figurative motion as if it were a resulting state (cf. He kicked the guitar to pieces).

Elaborating on Goldberg’s view, Peña (2008, pp. 1054–1055; 2009, p. 758) has further postulated the activity of the metaphor A PROCESS IS MOTION ALONG A PATH in AP resultatives. By virtue of their iterative component, AP resultatives such as He hammered the metal flat, He painted the house green, etc. involve (either instigated or non-instigated) processes of becoming. Peña maintains that such processes respond to the logic of Johnson’s (Reference Johnson1987) MOTION-ALONG-A-PATH image schema. They can thus be abstractly depicted as paths whose final point is the result of the action, i.e., the course of action whereby the metal becomes flatter and flatter is argued to entail metaphorical motion. Accordingly, in this metaphor, the beginning of the process maps onto the source of the path, the action(s) performed are the steps taken when traversing or moving along the path, and finally, the result maps onto the destination or goal. Processes have terminal points since “the metaphorical steps taken necessarily lead to a destination” (Peña, Reference Peña2008, p. 1055). In addition, processes involve changes that take place over time (Van Valin & LaPolla, Reference Van Valin and LaPolla1997, p. 83). Consider, however, resultatives like The knee to the head knocked him silly (COCA, 2003), I sneezed and the entire zipper of my dress burst wide open (COCA, 2008), I took a deep breath and tore it open (COCA, 2010), or Sailor finishes his beer, stands it on the ground and steps on it, crushing it flat (COCA, 1990), all of which depict processes of becoming. Now take, for the sake of illustration, the first example, The knee to the head knocked him silly. In the same way as ‘hammering a metal flat’ or ‘kicking someone black and blue’, this example is a case of the AP resultative. As such, it should in principle display the features described above in order for the metaphor A PROCESS IS MOTION ALONG A PATH to be operational. While The knee to the head knocked him silly has a terminal point, it does not abide by the condition that there must be some intermediate steps that gradually culminate in a change of state. Instead, the action is performed in such a way that it causes the patient to suddenly become silly. Such immediacy thus blocks the alleged perception of the event as a path, unless a PP is explicitly employed, as in He broke the vase into a thousand pieces or The life inside me kicked me into motion (COCA, 1997).

It follows from this that just because AP resultatives have a final point and generally describe repeated actions does not mean that the achievement of the result has to be construed as a path involving figurative motion, especially in the absence of motion markers such as to, into, out of, etc. The AP resultative is thus not a metaphorically interpreted goal phrase, as Goldberg (1995, p. 84) and Peña (Reference Peña2009) put it, but just a goal resultative phrase, as opposed to PP resultatives where the resultative element is necessarily metaphorical (cf. I have kicked it into a dozen pieces).

In closing, consider the following statement by Goldberg (1995, pp. 85–86, emphasis in the original):

Goldberg illustrates the fact that the verb break can collocate with a literal directional through the example He broke the walnuts into the bowl. It is clear that Goldberg’s example of the CMC does not display any kind of metaphorical understanding, i.e., the agent breaks the walnuts and as a result, they literally fall into the bowl. We have, however, distinguished between the CMC and the PP resultative since it is the latter that makes use of metaphorical motion in order to express a change of state, as opposed to a mere change of location. Break can thus occur in the PP resultative as attested by the following sentences: The transparent gel thinned and broke up into hundreds of skull-sized bubbles (COCA, 1999), She broke bread into pieces (COCA, 1996), They took the entire human genome and they broke it up into millions and millions of small pieces (COCA, 2004), to give but a few. Pace Goldberg’s claim, all these examples are motivated by the metaphor A CHANGE OF STATE IS A CHANGE OF LOCATION, in which the state of being in pieces is seen in terms of the destination of metaphorical motion.

As previously advanced, construals such as these in which motion towards a location interacts with a change of state are a consequence of conceptual conflation (see Hampe, Reference Hampe, Handl and Schmid2010, pp. 199–200). In line with Lakoff and Johnson (Reference Lakoff and Johnson1980, Reference Lakoff and Johnson1999) and Grady (Reference Grady1997), conflation is a conceptual mechanism through which speakers pair subjective experience with sensorimotor experience thus giving rise to primary metaphors like MORE IS UP (e.g., Gas prices are rising, based on the experience of seeing levels rise as more liquid is added), AFFECTION IS WARMTH (e.g., They greeted me warmly, based on the experience of feeling a person’s body temperature when being intimate), CATEGORIES ARE CONTAINERS (e.g., She broke us into groups, based on our experience of classifying objects into groups spatially placed in separate bounded regions), etc. In the same way, the metaphors STATES ARE LOCATIONS and A CHANGE OF STATE IS A CHANGE OF LOCATION are based on the conceptual conflation through experiential co-occurrence, which obtains, for instance, when in the shade of a tree we may feel alleviated from the effects of sunshine on us. Getting into the shade involves not only a change of location but also an accompanying change of state. Hence, A CHANGE OF STATE IS A CHANGE OF LOCATION licenses resultatives like the ones given above or many others of the type I drank myself into a stupor / a coma / an early grave. In them, the property acquired or the state reached by the patient is seen as the destination of figurative motion. This type of rationale, however, does not apply to AP resultatives such as Sam kicked Bill black and blue, given that, as has been argued above, becoming bruised cannot be a consequence of some sort of figurative motion.

We thus conclude that AP RCs do not inherit from the CMC. Rather, there are reasons to consider the CMC as a useful resource to express changes of state in two ways: one in which a motion verb and a motion PP are jointly used figuratively (e.g., She pushed me into despair); the other, in which a change-of-state verb is combined with a figurative motion PP denoting a resultant state (e.g., He melted the butter to liquid). In the second case, the verbal predicate melt, which is not a caused-motion verb, is coerced into the CMC framework.

We finally move on to the analysis of a selection of examples, with special focus on motion.

4. Motivating motion

In Section 1 we briefly noted the importance of specifying the underlying mechanisms that motivate lexical–constructional integration. It goes without saying that this specification is necessary in order to fully determine, in the most economical way, the conditions under which lexical items are incorporated into a given construction. So far, in Sections 2 and 3, we have concentrated on the high-level metaphor A CHANGE OF STATE IS A CHANGE OF LOCATION due to its ubiquitous presence in RCs with resultative phrases like to pieces, into bits, into despair, to death, among many others. But there are other productive high-level metaphors and metonymies at work, which also deserve discussion. We will now address these other mechanisms. In order to do so, we have organized this final section as follows. Subsection 4.1 provides the reader with a necessarily brief theoretical overview of high-level metaphor and metonymy. In order to illustrate the fundamental licensing role played by these abstract cognitive operations, we zoom in on the issue of (figurative and literal) motion in the two broad conceptual categories under scrutiny here, i.e., changes of state and changes of location. Then 4.1.1 deals with the verbal predicate run within the AP resultative, as well as some additional cases of the PP resultative. Finally, 4.1.2 is entirely devoted to the CMC.