2.1 Predicates of physical disturbance

The predicates in (1) and (2) above involve a “physical disturbance” in an affected object: an irregularity in the object that is part of the object, and may be imagined to move around the object, also known as the “host” of the disturbance (Karmo 1977).

These predicates are distinguished by the availability of a count noun form for the root, so that both verbal and adjectival forms may be paraphrased with a nominal predicate as shown in (3).

This contrasts with other deverbal adjectives of CoS verbs, at least, those that describe physical damage to inanimate objects.³ For instance, damaged in (4) has a root with a nominal form, but only as a mass noun, while the root of shattered has no nominal correspondent at all (5).

(6) below lists – not exhaustively – other examples of physical disturbance CoS verbs.

Among these, crack has been the most discussed, and will feature most in the following discussion. Examples with dented and scratched are generally provided for a wider perspective, with occasional support from crease. For brevity, examples with other predicates have not been systematically included.

2.2 Background: Shared scalar properties in related adjectives and CoS verbs

Change of state predication can be said to be linked to some kind of stative component, or stative core, based on which the change is described. This link is classically captured by the use of a become predicate, used in the Generative Semantics literature (see e.g. Katz 1970), and given more formal development in Dowty (1979). With recent work on scalar structure (Kennedy 1997; Kennedy & McNally 2005), the nature of the scale associated with the stative core has been argued to predict the telicity and interpretational properties of CoS verbs (see e.g. Hay et al. 1999; Kearns 2007; Kennedy & Levin 2008). For instance, for an open scale adjective such as wide (7), the corresponding deadjectival CoS verb widen cannot be telic: it is compatible with a for X time-adverbial but not with an in X time adverbial (7a).⁴ The progressive form of the CoS verb entails the perfective (7b), and both CoS verb and adjective are incompatible with the adverb “completely”.

In contrast, a CoS verb derived from an adjective with a maximum value on its scale can be telic, and compatible with in X time but less so with for X time (8a). For these CoS verbs, the progressive does not entail the perfective (8b), and both adjective and the related CoS verb are compatible with “completely” (8c).

The scalar correlation between adjective and corresponding deadjectival CoS suggests that a parallel correlation could also hold in the other direction, i.e. between CoS verbs and their related deverbal adjectives. This is the assumption in Rappaport Hovav (2014), who points out that underived CoS verbs also show the range of scale types that have been observed for adjectives. CoS verbs may lexicalize scales that are open or closed on one or both ends (Kennedy & McNally 2005), the scales may be so-called two-point, or multi-point scales (Beavers 2008), the former describing punctual events, and the latter associated with gradual change. In each case, the adjective is assumed to correspond to the stative core, or “property base” of the CoS verb.

The scale-based classification of related verbs and adjectives in Table 1 (Rappaport Hovav 2014: 266) provides a uniform treatment for morphologically related verbs and adjectives, irrespective of whether the verb or the adjective is morphologically basic.

Looking more closely at the members of Table 1, the pairs die/dead and crack/cracked are assumed to be analogous in scale structure. Both CoS verb and adjective are taken to be associated with a two-point scale, although in the former, the adjective is presumably basic, while in the latter, the CoS verb would be basic. Interestingly, in their discussion of adjectival scale structure, Rotstein & Winter (2004) categorize cracked as a “partial” adjective. In boundedness terms, cracked would be associated with a lower-bounded scale that is open on the upper end. The discrepancy in these two classifications, albeit made in passing, suggest the need for a closer look at these predicates. The next section will show that, differing from these previous assumptions, deverbal cracked is gradable, with both a minimum and a maximum value.

2.3 Gradability in disturbance predicates

This section examines the event and scalar structural properties of crack and cracked, arguing that they differ from die/dead, contrary to the assumption in Table 1 that both pairs are associated with two-point scales. The classification is appropriate for die, which is telic, and incompatible with for X time adverbials. The punctual nature of die is mirrored by its non-gradable related adjective dead: #more dead is not felicitous (Kennedy & McNally 2005; Beavers 2008).⁶

(9) below suggests that crack can indeed be treated as a punctual CoS verb, applying diagnostics from Beavers (2008), following Kearns (2000: 206).

With an adverbial such as in five minutes, crack allows only the reading where the cracking event took place after five minutes (9). This much is analogous to die, but the kind of event described by crack actually allows for more than one temporal structure. A cracking event description may also be compatible with durative for adverbials (10a)–(10b). Consistent with the possibility of the verb’s durative use, the deverbal adjective cracked allows a comparative use with more (10c).⁷

In keeping with the parallel between durative verb and gradable adjective illustrated in (10a)–(10c), the comparative use of deverbal adjectives of the cracked type is consistently attested across other kinds of physical disturbance deverbal adjectives, as shown by the examples in (11), all found online.

These examples are surprising if the verb and adjective can only be associated with a two-point scale. Imaginably, the contrast between die/dead and crack/cracked could be that the former pertains to (formerly) animate entities, whereas the latter describes physical objects. But there are CoS verb/deverbal adjective pairs that describe physical objects, but like dead, are non-gradable. Like both die and crack, the (intransitive) verb shatter describes a punctual event: the in X time adverbial in (12a) receives only a reading where the shattering event occurred after the stated amount of time. Unlike crack, and like die, however, deverbal shattered is not gradable (12b).

The parallel between die/dead and shatter/shattered shows that the atelicity of the verb crack and the gradability of cracked and other deverbal adjectives of physical disturbance is particular to the class of physical disturbance predicates itself, and not merely a consequence of predicates applying to inanimate physical objects.

2.4 Physical disturbances are associated with a closed scale

The scale structure of physical disturbance predicates bears further scrutiny as it finds no straightforward parallel with predicates of other scale types discussed in the previous literature. The availability of durative and gradable readings notwithstanding, the assumption in Rappaport Hovav (2014) that crack is a punctual verb associated with a two-point scale does seem consistent with the interpretation of examples such as (9) above, repeated here.

One might compare crack with degree achievements such as cool, which allow both telic and atelic verbal uses (Dowty 1979; Abusch 1986; Kennedy & Levin 2008), as in (13), but crack is not clearly analogous to them. The telic and atelic uses of cool are asymmetrical: the telic (13a) entails that the soup became cool, but the atelic (13b) does not.

In contrast, whether something cracked in X time (14a) or cracked for X time (14b), it is entailed to have cracked.

Alternatively, one might compare crack to fill or empty, which are presumably associated with scales that are closed on both the lower and upper ends (Kennedy & McNally 2005). This parallel seems supported by adverbial modification facts for crack, which is compatible with the adverbs completely (15a) and partially (15b), suggesting a totally closed scale.

Nonetheless, crack still differs from fill/empty in taking adverbials such as slightly (16), which generally indicate a lower rather than an upper bound (Rotstein & Winter 2004).

Moreover, whether with completely or slightly, verbal cracked entails the situation described by its corresponding adjectival predicate (17a). In contrast, to the extent that fill slightly is acceptable, it does not entail The basin is filled. Adjectival filled is entailed only by fill completely.

Setting aside for now the overlaps and contrasts with other classes of predicates, it seems reasonable overall to classify physical disturbance CoS verbs as being associated with a gradable scale that is closed on both ends. This understanding of the scale for disturbance CoS verbs receives support from the scalar properties of their related deverbal adjectives, as evidenced in degree modification using the adverbials badly and well. Deverbal adjectives of physical disturbance seem to combine most naturally with badly (18) to indicate the degree of the disturbance.

This use of badly is synonymous with the use of well – also indicating high degree – as the examples in (19) show.

Kennedy & McNally (1999) and McNally & Kennedy (2002; 2013) propose that the degree modifier use of badly and well is possible only with adjectives that (i) are gradable, and (ii) associated with a totally closed scale.⁸ We saw earlier in (10)–(11) that physical disturbance adjectives are indeed gradable. The examples below show that disturbance adjectives occur felicitously with so-called “endpoint-oriented modifiers” (McNally & Kennedy 2002: 2) such as completely (20) and partially (21), corroborating the claim that they are associated with a scale that is closed at both ends.

Overall then, deverbal disturbance adjectives should be understood as gradable, closed-scale items, contrasting with non-gradable dead. Both these adjectives and their related CoS verbs should be associated with multi-point scales that are closed on both ends.

3.1 Different dimensions in physical disturbance adjectives

Deverbal disturbance adjectives allow for different, but non-contradictory, interpretations in certain contexts. In particular, when modified by the adverb badly, as in (22), these adjectives yield at least two interpretations. The examples in (22) may be understood to describe (i) many scratches in the print, or many dents in the car (quantity), or (ii) few, perhaps even only one scratch or dent, but the relevant scratch or dent is of a serious kind: say, very long or very deep (quality).

Of course, these readings are not mutually exclusive, and indeed, the use of badly could indicate a high degree of both and possibly more.

Nonetheless, in many examples, it is possible to distinguish one particular reading being expressed by a deverbal disturbance adjective, as suggested by (23)–(25) below. My interpretation of each example is given in parentheses after the relevant sentence.¹⁰

I did not manage to find clear examples of badly/well scratched describing a single deep scratch, perhaps reflecting real-world situations of scratched surfaces described as “badly scratched”, which are likely to involve multiple scratches. There were, however, numerous illustrative examples of both interpretations of badly creased. I have selected examples which were accompanied by a photograph picturing the intended interpretation.

In what follows, I explore the interpretations introduced above, illustrating the multidimensional nature of deverbal disturbance adjectives to eventually build up a picture of how the scalar structure of disturbance predicates comes about.

3.2 Background: sources of multidimensionality

Multidimensionality in adjective meaning can be considered to exemplify a kind of polysemy involving so-called “non-antagonistic”, or “cooperative” readings (Cruse 1995: 44). When cooperative readings are involved, apparently contradictory propositions can be acceptable at the same time. Cruse discusses cooperative readings – known as facets – that certain nouns exhibit. For instance, the noun book has a [TEXT] and [TOME] facet, and in answer to the question Do you like the book?, both yes and no could be truthful answers, one pertaining to a different facet ([TEXT]: yes, it’s very well-written; [TOME]: no, the paper is of poor quality and the text is too cramped). Somewhat analogously, adjectives may be subjectively interpreted, allowing for “faultless disagreement” (Kölbel 2004; Lasersohn 2005) in which apparently contradictory propositions may be truthfully uttered by different speakers.

Subjectivity in adjectives has been attributed to at least two major factors: judge-dependency, which could arise in more ways than one (Kennedy 2013), possibly reflecting how speakers negotiate the discourse (Barker 2013), and, more relevant for the current discussion, multidimensionality (Sassoon 2013). Sassoon (2013) differentiates multidimensional adjectives such as healthy, which can combine with respect PPs, e.g. healthy with respect to cholesterol, from (uni)dimensional adjectives such as long, which may describe different kinds of dimensions, but only access one in a given context. Single dimension adjectives are incompatible with respect PPs: #The wedding/table is long with respect to temporal duration (but not space) (Sassoon 2013: 338 (2c, d)).

Solt (2018) further links multidimensionality to the compatibility of adjectives with “ordering subjectivity”, noting that dimensional adjectives such as tall allow faultless disagreement only in non-comparative contexts, and not in a comparative. That is, two speakers could disagree about whether someone was tall without one of them being necessarily wrong, but in a disagreement about which of two individuals is taller, one of them must be wrong: i.e., no faultless disagreement is possible in this case. In contrast, subjective adjectives such as tasty and beautiful allow for faultless disagreement in both simple and comparative contexts. On this basis, Solt distinguishes between adjectives that are objective, subjective, or both, in their comparative forms, reflecting disagreement that would be fact-based, opinion-based, or both. Of these, factual disagreement is attributed to multidimensionality and opinion-based disagreement to judge-dependency. Not surprisingly, adjective interpretation can be influenced by both kinds of factors.

The rest of this section discusses objective and subjective judgements in the use of deverbal disturbance adjectives, showing that subjective judgements arise only in degree modification and comparative contexts.

3.2.2 Subjectivity in physical disturbance deverbal adjectives

In contrast to simple predication contexts, deverbal physical disturbance adjectives show faultless disagreement in comparative contexts. As the following examples suggest, the subjectivity in these situations is attributable to multidimensionality. Individual dimensions such as quality or quantity can be independently specified using a respect PP (28a). The different dimensions do not require a comparative context per se to surface: degree modification is sufficient for a dimension to be individuated (28b), or quantified over (28c).

1. (28)
1. a.

1. Your front bumper is more dented with respect to dent size, but your rear bumper is more dented with respect to the sheer number of dents.

1.  
1. b.

1. Your front bumper is badly dented at least with respect to dent size/number of dents.

1.  
1. c.

1. Your front bumper is badly dented in every respect you can name!

A small online survey using only the adjective cracked with 13 respondents,¹¹ revealed that speakers may evaluate how badly cracked a host entity is relative to another with reference to a variety of dimensions. Respondents saw pictures of pairs of the same item, e.g., a picture of a cell phone, each showing a different pattern of cracks, as exemplified in Figure 1.

Figure 1: A sample comparison pair of cracked cellphone images from the online survey.

For each pair, respondents were asked to determine which item (say, Phone A or Phone B) was more badly cracked. They were given four options from which one should be chosen: (i) Phone A is more badly cracked; (ii) Phone B is more badly cracked; (iii) Phone A and Phone B are equally badly cracked; and (iv) none of the above. Each question (with a total of eight questions) was followed by a request for comments on why the respondent had made a particular choice. The responses showed variation in how relative damage was evaluated – while there was often some consensus as to which item (or neither) was more badly cracked, responses were by no means unanimous. Consensus ranged from cases of 80% agreement on one multiple choice option, to cases in which speakers were more or less evenly split, with approximately 40% of converging on one option and another 40% on another option.¹² Comments showed that respondents considered not only perceived severity of cracks and number of cracks, but also the position of the cracks on the host entity, whether the functionality of the host entity is perceived to be more or less affected, how the cracks are arrayed, and potentially other factors. As (29) below illustrates, each factor mentioned above is specifiable using a respect PP, although the example is unsurprisingly wordy.

1. (29)

1. This mirror is more badly cracked than that one with respect to the position of the cracks/how badly its functionality is affected/how the cracks are arrayed on the surface …

Although small, the survey results support the understanding of physical disturbance deverbal adjectives as multidimensional. While judge dependence is not at issue in the current discussion, the survey suggests that this factor is also relevant in evaluations of degree of damage.

3.3 Evoking a single dimension: the case of much

In the contexts discussed above, the different dimensions of physical disturbance deverbal adjectives are distinguishable interpretively and via respect phrase specification. This subsection provides distributional evidence from adverbial modification with much, to further illustrate that individual dimensions of these adjectives may be accessed independently of each other.

As the examples in (30) show, the use of much with deverbal adjectives of physical disturbance evoke only the “quantity” dimension: these examples can only be understood as describing many disturbances on a surface. The restricted interpretation is robust across different adjectives, and different objects of description.

The compatibility of much with deverbal disturbance adjectives is somewhat unexpected, as much is not clearly known to be compatible with totally closed scale adjectives. As suggested in Kennedy & McNally (2005), much generally seems to occur with lower closed scale (deverbal) adjectives that are not upper closed, as demonstrated by the behaviour of desired (31a), which entails minimal standards (31b) (i.e., “not desired” means “totally not desired”), and is not compatible with modifiers indicative of maximal standards (31c).¹³

Yet deverbal disturbance adjectives also show traits of total closed-scale adjectives in allowing degree modification via badly, and in being compatible with adverbs such as completely (20) that target the maximal degree of a scale. The conflict cannot be avoided simply by broadening the purview of much to include totally closed-scale adjectives, since this does not account for why the quality dimension is not compatible with much.

This suggests disturbance adjectives are compatible with both totally closed and non-upper closed interpretations, an apparently contradictory situation. Below, I discuss how these two interpretations can be resolved.

3.4 Assembling the deverbal disturbance scale

In this section, I propose that the scalar properties of deverbal physical disturbance adjectives are related to the physical properties of the entity they describe. Based on the preceding discussion, there are three components to be accounted for: the lower bound, the gradable interval, and the upper bound. I discuss each briefly below.

The source for the lower bound seems straightforward enough. This would be the physical instantiation of the relevant disturbance, so that the host object could no longer be described as unblemished. While there could be faultless disagreement about how exactly to describe the disturbance, e.g. scratched as opposed to scuffed, dented as opposed to dinged, etc., it is clear these adjectives entail a physically instantiated lower bound.

The gradability of the adjective would come from more than one dimension. As discussed in Section 3.2.2, these could include the quantity of disturbances, the qualitative physical properties of the disturbance, such as its depth and spatial extent, and also factors such as the positioning of the disturbance, e.g. how centrally it is perceived to be located relative to the host object, how multiple disturbances are arrayed with respect to one another, e.g. some number of isolated cracks, or the same number of cracks forming a web, etc. Certain factors involve a level of judge dependency, such as how much the observer perceives functionality to be affected, among other criteria (McNally 2011; Foppolo & Panzeri 2013). It seems reasonable to assume that these dimensions themselves do not impose an upper or lower bound, excepting the quantity dimension, the lower bound of which is presumably tied to the lower bound from the existence of the disturbance.

The upper bound of the disturbance adjective scale would come from the spatial extent of the host object. Something is completely cracked/dented/scratched etc. presumably when all its (relevant) parts are cracked or dented or scratched etc. That the entity to which an adjectival description applies provides an upper spatial bound for the adjective is perhaps true for many adjectives, and is not necessarily unique to physical disturbance adjectives. This is true for adjectives with an upper-closed scale, e.g. dark, or those with a lower-closed scale, e.g. dirty. For instance, for The sky is completely dark to be true, all parts of the sky have to be completely dark (Kennedy & Levin 2008). Similarly, something can be completely dirty if all its parts are covered in dirt, although the scale associated with dirty is presumably open on the upper end.

Deverbal disturbance adjectives are, however, distinct from both dark and dirty. For adjectives such as dark, which define a maximal degree not based on the spatial extent of the theme, evaluating the truth of something being completely dark actually appeals to two maxima: the attainment of maximum darkness, and the maximum darkness of every part of the theme. For a lower-closed scale adjective such as dirty, completely applies only to the spatial dimension, and crucially, does not suggest there is some maximum degree of dirtiness. Something could be completely dirty in being covered with dirt, but more dirt could get onto it (32).

In contrast, the spatial dimension features directly into the upper bound of deverbal physical disturbance adjectives. This is presumably because physical disturbances require that an entity maintain physical integrity to host a disturbance. For instance, there is a point where cracking will lead to breaking, in which case the maximum degree of cracking will have been exceeded, as the constructed (33) attempts to evoke.

While this maximum is more difficult to determine for scratching and denting, there are examples suggesting they similarly involve a loss of structural integrity to the host object. Although I could not find examples with the adjectival forms of these predicates, the naturally-occurring verbal and nominal examples below support the notion that structural integrity of the host object provides an upper bound to the degree of cracking, scratching, and denting. (34a) describes a pump losing its structural integrity from excessive cracking, splintering, and denting to become a collection of metal and wood. Online searches for scratched often bring up discussions of itchy skin, for which there may be no scratch marks. But (34b), involving nominal scratch in a description involving photographic film, suggests tearing as a likely result of breaching the upper bound of “scratchedness”.

As the preceding examples suggest, either a sufficiently large number of disturbances, or a sufficiently serious disturbance, could exceed the spatial limits imposed by the host entity.

This section has demonstrated the multidimensionality of deverbal physical disturbance adjectives, illustrating the distinct scalar properties of each dimension, and their origins. The next section explores whether their related CoS verbs also allow linguistic access to these dimensions.

4.1 Disturbance CoS verbs allow the same adverbs as their related adjectives

In terms of scale structure, the generally assumed correlation between related verbs and adjectives seems to apply also to disturbance CoS verbs. These verbs allow modification with completely (36) and with badly (37) on its degree interpretation, as do their adjectival counterparts.

Despite the shared compatibility in adverbial modification, physical disturbance CoS verbs do not show the same access to the different dimensions of the disturbance and its host entity as their related adjectives do. Below, I show that disturbance CoS verbs are also multidimensional, but only in a conceptual sense.

4.2 Disturbance CoS verbs are conceptually multidimensional

This subsection discusses attested uses of disturbance CoS verbs modified by badly. I show that these examples allow interpretational distinctions between different dimensions of the physical disturbance event, but unlike for adjectival predicates, these dimensions are not easily specifiable with respect phrases. Solt (2018) notes that multidimensionality for certain adjectives is more grammatically available, with component dimensions relatively individuable, specifiable, and quantifiable, via respect phrases, while in other cases, the component dimensions are available to conceptual interpretation but may not allow grammatical access. Building on this observation, Ruiz & Faroldi (2022) distinguish between “quantificational” and “conceptual” multidimensionality, which I suggest below is how physical disturbance adjectives differ from their related CoS verb.¹⁴ That is, modification of disturbance CoS verbs by badly may be interpreted as damage to a serious extent, coverage of a proportionally large area, a combination of both, or possibly other manifestations, but these interpretations arise from a combination of context and world knowledge, and are not easily accessible using “respect” PPs and quantification. The examples below illustrate that disturbance CoS verbs may be used to identify individual dimensions, even though most of the time the intended dimension is not straightforwardly distinguishable. The component dimensions, however, do not seem to be easily specifiable via respect phrases.

Examples of cracked badly from a Google search generally tend to suggest at least a deep crack, if not cracking that is both deep and extensive. Where photos accompanied the text, most cases involved cracking that was both deep and extensive over some part of the surface described. Several examples involved one or two deep cracks, as in (38a), from a YouTube video, where the text accompanied video of a miniature concrete drum mixer with a single crack. I managed, however, to find one clear example in which verbal cracked badly was accompanied by a photograph showing the stripes in question to be covered in cracks (38b).

Similar options are possible for scratched (NP) badly. Clear cases of intransitive disturbance CoS verbs with badly are difficult to find. I have thus used examples with a transitive verb where it is clear that CoS is being expressed.¹⁵ (39a) shows a clear example where something said to be scratched badly was affected by a single deep scratch. In contrast, (39b) is accompanied by a photo that showed multiple scratches over some part of a car door, suggesting that damaged surface area also factors into the verbal use of scratch.

Although these interpretations are conceptually available, disturbance CoS verbs are less compatible with respect PPs than their adjectival counterparts. The judgements reported in this section are from an informal survey of seven native speakers, almost all of whom (6/7) found (40a) straightforwardly acceptable (one found it degraded), while judgements for (40b) varied across speakers (1/7 found it acceptable, 3/7 found it degraded, 3/7 found it unacceptable).

The asymmetry between disturbance adjectives and verbs in their compatibility with respect phrases is maintained, albeit to a lesser extent, in quantificational contexts (41). The judgements are subtle, but consultants generally preferred adjectival examples such as (41a) and (42a) to their verbal counterparts such as (41b) and (42b). While (41a) was degraded for some speakers (3/7), some found it fully acceptable (4/7). In contrast, no speaker found (41b) fully acceptable.

All speakers found (42a) fully acceptable. In contrast, judgements were mixed for (42b): 2/7 found it acceptable, with others finding it either fully (1/7) or less than fully unacceptable (4/7).

Causative uses of the CoS verb, such as (43), were generally judged to be acceptable with a respect phrase, with only one speaker finding the examples degraded.

This asymmetry is not because respect PPs are unable to modify verbs. Clearly multidimensional verbs, whether stative (44a) or CoS (44b), occur felicitously with respect PPs.

The reduced compatibility with respect PPs holds on the one hand between (at least intransitive) disturbance CoS verbs and their related adjectives, as well as between disturbance CoS verbs and other verbs, both stative and CoS. The contrasts suggest that the multidimensionality of physical disturbance CoS verbs is not necessarily grammatically represented by its stative core, or at least, not represented in the same way as for their related adjectives. Of course, there remains the question that respect phrases with disturbance CoS verbs may be more acceptable in certain contexts. I discuss these questions in the next section.

5.1 Physical disturbance deverbal adjectives do not entail a preceding change

Morphologically speaking, physical disturbance adjectives show the participial forms of their related CoS verbs, and can be considered to be derived from the verb. Semantically, however, they do not necessarily express the resultant state of an event of change described by the corresponding CoS verb. The following examples show uses of physical disturbance deverbal adjectives that describe cracked, dented, and scratched surfaces that have clearly not undergone any cracking, denting, or scratching. The cracked surface of the pumpkin in (45a), shown in an accompanying photograph online (accessible from the url given), simply describes how the pumpkin skin is textured. The dented surface in (45b) is for a visual model of a helmet – the effect is achieved through computer modelling. (45c) reproduces an online product description for a phone vinyl decal, and the scratched surface has clearly not been scratched, but is rather from the graphical representation of scratch-like vertical lines.

This independence of the adjectival description of physical disturbance from a preceding event is consistent with prior studies of deverbal adjectives morphologically related to other CoS verbs such as bend/bent, darken/darkened etc., for which it has been argued that these predicates may describe only spatial change that does not involve temporally prior events of change (Gawron 2009; Koontz-Garboden 2011).

Another possibility is that disturbance adjectives could be denominal, the -ed suffix analogous to that found in nominal expressions such as a bearded lady or a flowered dress (Koontz-Garboden 2012; Nevins & Myler 2014). This is not likely to be the case for physical disturbance adjectives, which, at least morphologically, should be considered deverbal. This point finds support in derived adjectives with root-identical verb and noun forms, but which clearly show the past participle form of the verb, such as broken and torn. Both have corresponding basic verb and noun forms break and tear, but the adjective is clearly deverbal, and not *breaked and *teared as would be expected on a denominal analysis.

The relationship between disturbance verbs and nouns, which are form-identical, cannot be properly addressed here, given that the focus of this work has been on verbs and adjectives. In dictionary entries, the verb is typically listed as the etymologically prior item, but what factors in general allow verbs and nouns to share the same form, and how their interpretations may be related, is itself a question for further investigation (see e.g. Croft 2001: ch.2).

5.2 Differing multidimensionality in physical disturbance predicates

This section discusses possible meaning representations for physical disturbance adjectives and verbs, with a view towards reflecting their different manifestations of multidimensionality.

Representing the multidimensionality of physical disturbance adjectives finds a useful model in Solt (2018). Solt defines a scale S as a triple {D, ≻, dim}, with D a set of degrees, dim a dimension of measurement, and ≻ an ordering relation on D (Solt 2018: 78 (34)). Adjective meaning is underspecified, being relativized to a context c, comprising world, time, and judge so that gradable adjectives “lexicalize dimensions rather than particular scales or measure functions” (p.79). A gradable adjective would describe a situation in which some property is true of an individual to some standard degree or higher, relative to a dimension in some context c (46).

For multidimensional adjectives, ${\mu }_{\mathit{\text{DIM}}}^c$ integrates various component dimensions ${\mu }_{\mathrm{DI}{M}_1^c}^c$, ${\mu }_{{\mathit{\text{DIM}}}_2^c}^c$, …${\mu }_{{\mathit{\text{DIM}}}_n^c}^c$ of the adjective (Solt 2018: 81). Following this approach, a physical disturbance adjective such as cracked would describe the degree d to which its host item x is cracked in context c. The measure function (47a) that yields the degree of cracking could be given the meaning in (47b) (based on Solt’s representation of the adjective dirty), making reference to the extent of the disturbance based on dimensions relevant in the context, relative to the spatial extent of the host entity.

Physical disturbance CoS verbs would access the same scale as their corresponding adjective. As the preceding section suggests however, the multidimensionality of the scale is conceptually available but not directly accessible to grammatical operations with respect phrases and quantification. This suggests a measure of change function (Kennedy & Levin 2008) representation might not be appropriate for these verbs. Below, I discuss a possible meaning for physical disturbance CoS verbs, and speculate on why individual dimensions for these verbs may be less accessible.

A CoS verb describes an event of change, and for scalar predicates, a corresponding change in degree along the relevant scale. In the semantic representation for deadjectival CoS verbs in Beavers et al. (2022), based on work in Beavers & Koontz Garboden (2020), separate provisions are made for the event of change and its associated degree of scalar change. The representation makes use of the become operator which applies to a state denoted by the root, and includes an open difference degree that is calculated from an initial degree and a final degree on the scale evoked by the root. Loosely adapting their approach for current purposes, a possible meaning for CoS cracked could be as in (48) below.

In (48), the become component indicates the CoS status of the predicate. The state argument (i.e., s) of become is associated with the root, which provides a verbal standard for the degree of cracking that CoS crack requires. That is, CoS crack expresses an event of change to a state of something being cracked to a degree that is at or above what is standardly considered to be cracked. In addition, CoS crack makes reference to a difference degree d_d. This difference degree is obtained from the difference between the degree to which the theme of crack is cracked, relative to some dimension, at the end of the cracking event (d_f), and the degree to which it is cracked, again, relative to some dimension, at the beginning of the cracking event (d_i), which is required to be smaller than d_f.¹⁶

Given this representation, a physical disturbance CoS verb fundamentally describes an event, but this event is associated with a degree of change that is in turn associated with one or more dimensions. I suggest that the dimensions of the change are not fully accessible to quantification or direct reference because multidimensionality only comes into play for determining the degree of change, unlike for the corresponding disturbance adjective meaning, where a measure function comprises the main component.¹⁷

For physical disturbance CoS verbs, calculating degree of change introduces a further element of complexity – in the case where an item was already damaged, say, a windshield with a pre-existing inch-long crack, which would perhaps not be considered badly cracked. Now suppose the windshield cracks further with age, and the original crack extends another inch. Imaginably, with a two-inch crack, the windshield is badly cracked, but did it crack badly? This kind of scenario suggests that assessing the difference degree for a physical disturbance CoS verb involves more complexity, and may require knowledge of the history of the host, whereas history is not relevant for physical disturbance adjectives. Potentially, factors such as these render individual dimensions less accessible to the speaker/hearer with CoS verbs, although the multidimensional nature of a physical disturbance would remain conceptually available, and thus respect phrases are less acceptable but not fully ruled out for CoS verbs.

A remaining question is why disturbance CoS verbs seem more acceptable in comparison contexts, or when used causatively. I am unable to resolve this question, but I speculate that the improved acceptability of examples such as (42b) and (43) is pragmatic in nature. The comparative context could be more acceptable than the verbal predicate alone because the verbal predicate describes an event, but the comparative compares degrees of change (see e.g. the discussion of more than in Beavers et al. 2022). This meaning of the comparative could be a factor in facilitating the reference to individual dimensions. Mention of the cause of the disturbance could draw attention to the process of sustaining damage, and the respect phrase could perhaps be interpreted as indirectly describing the kind of force exerted. Whatever the reasons, deverbal disturbance adjectives clearly require no contextualizing for individual dimensions to be specified. That disturbance CoS verbs allow for this only in relatively richer contexts suggests the workings of pragmatic accommodation.