Speaking of death

Abstract

As a human-specific trait, language offers a unique window on human cognition. Grammatical constraints on the ways we speak about events, for instance, have long been thought to reveal the representational formats that our minds impose on the ways that we think about events. In recent research, verbs that name events of death have stood out as key counterexamples to standard theories of the grammatical constraints on possible verbs. The special status of these thanatological verbs raises two important questions: why, given the vast number of verbs in any language, is it that verbs of death hold this special status, and what do they tell us about the restrictions on the representational format for possible verbs? This paper reexamines the evidence coming from verbs of death, confirming that they are counterexamples to standard theories, but that their behaviour suggests a more revealing constraint on our mental representations—that our minds impose strict restrictions on the format of asserted meaning. Thus, the constraints on linguistic representation and the human mind offer a unique perspective on the mental representations of thanatological phenomena.

This article is part of the theme issue ‘Evolutionary thanatology: impacts of the dead on the living in humans and other animals’.

1. Introduction

Do the ways we speak about events of death reveal anything special about how we think about events of death? Studies on verbs and verb meanings have been seen as core to our understanding about how we mentally represent events. Decades of research have worked to isolate how different aspects of events are encoded in grammatically relevant meaning components across different parts of a sentence, with the verb acting as a key locus of event information. These studies have furthermore demonstrated that there are substantive constraints on what meanings can be packaged into a single verbal concept.

The main focus of this paper is how verbs can be separated into those that encode the manner in which an event process unfolds and those that encode the result of an event having unfolded, and what constraints exist between these two meaning components. Verbs like those in (1) encode the manner in which the process of an event unfolds, for instance as a process of something being screamed or someone being scrubbed or rocked. Other verbs like those in (2) encode the results of an event, for instance an event resulting in something being repeated or someone becoming clean or arriving. These encoded meanings are those that are lexically entailed by the verbal concept; that is, the verb scream always entails a manner (of screaming) while the verb repeat always entails a result (of something being repeated).

• (1)

  • a. Kevin screamed the lyrics.
  • b. Laura scrubbed Billy.
  • c. Willy was (gently) rocked.
• (2)

  • a. Kevin repeated the lyrics.
  • b. Laura cleaned Billy.
  • c. Willy arrived.

This categorization of verbs into manner versus result is one that cuts across our more naturally intuitive semantic verb classes. The verbs scream and repeat are drawn from a wider semantic class of verbs that name events of speech given in (3). The verbs scrub and clean are only a couple of the verbs that name events of changes of state given in (4). The verbs rock and arrive come from a class of verbs that name events of motion given in (5). Within each of these classes, we can identify subclasses of verbs that encode either the manner of the event, those in (a), or the result of the event, those in (b).

• (3) Speech Verbs

  • a. Manner: babble, bellow, chant, cry, grumble, holler, mumble, mutter, roar, scream, shout, sing, snap, stammer, whisper, yell …
  • b. Result: admit, announce, blurt, claim, convey, declare, mention, proclaim, propose, remark, repeat, report, say, state …
• (4) Change of State Verbs

  • a. Manner: bash, dust, hammer, hit, kick, polish, pound, pour, scour, rub, scrub, strike, sweep, tamp, tap, wash, wipe …
  • b. Result: break, chip, clean, crack, dissolve, divide, drain, empty, freeze, melt, pop, remove, rip, shatter, sink, splinter, tear …
• (5) Motion Verbs

  • a. Manner: crawl, dance, gallop, glide, jump, limp, meander, rock, roll, run, saunter, slither, spin, swim, swing, taxi, wade, walk …
  • b. Result: approach, arrive, ascend, come, depart, descend, enter, exit, fall, flee, go, increase, leave, near, recede, return, rise …

Thus, it seems that manner and result are basic building blocks of verb meanings apart from a verb's semantic field, and an investigation of these basic meaning components of verbal concepts may give us insight into our underlying mental representations for events. As such, if there is anything special about how we think about events of death, an examination of our verbal concepts for events of death in terms of manner and result might be a good place to start.

At first blush, the class of verbs we use to speak about typical events surrounding death seem no different from other verbal classes. There are verbs like those in (6) that encode an event process, for instance of shooting or stabbing, and there are others like those in (7) that encode an event result, such as in being killed or dying.

• (6)

  • a. Laura shot Billy.
  • b. Willy was (viciously) stabbed.
• (7)

  • a. Laura killed Billy.
  • b. Willy died.

Thus in both our more mundane everyday discussions of events and of those events that seem more critical to human preservation, a complementary pattern emerges. Verbs like those in (1) and (6) encode the manner of an event, directly telling us how an event unfolds, but not what the result of that event unfolding is. Other verbs like those in (2) and (7) encode the result of an event, directly telling us what the result of an event unfolding is, but not how that event unfolds.

To investigate these distinctions more systematically, linguistic research draws on naturally occurring examples found in corpora and native speaker judgements of the interpretations and acceptability of constructed examples. For example, native speakers permit sentences with manner verbs to occur with an additional predicate that encodes the result of the event, often expressed by an adjective or a prepositional phrase, as in (8), but reject similar sentences with result verbs, as in (9).¹

• (8)

  • a. Kevin screamed the lyrics to life.
  • b. Laura scrubbed Billy dirt-free.
  • c. Willy was rocked to sleep.
• (9)

  • a. *Kevin repeated the lyrics to life.
  • b. *Laura cleaned Billy dirt-free.
  • c. *Willy arrived to sleep.

This same pattern of native speaker judgements is attested with verbs related to death. Resultative predicates are permitted with manner verbs like shoot and stab in (10), but rejected with result verbs like kill and die in (11).

• (10)

  • a. Laura shot Billy dead.
  • b. Willy was stabbed to death.
• (11)

  • a. *Laura killed Billy dead.
  • b. *Willy died to death.

This complementary pattern between manner and result verbs has been attested again and again across a variety of languages, suggesting that the mental representation of any verb in any natural language is formatted such that it encodes only a manner meaning or a result meaning. This constraint on the representational format of our verbal concepts, dubbed manner/result complementarity as given in (12), has been proposed and defended by Levin & Rappaport Hovav [1–3]. In this paper, I take manner/result complementarity to be an empirically testable claim about the limits of what meanings we can bundle together and store in our mental lexicon for any single verbal concept.

• (12) Manner/result complementarity: Manner and result meaning components are in complementary distribution: a verb lexicalizes only one. Levin & Rappaport Hovav [4, p. 50]

Investigating potential counterexamples to manner/result complementarity has been important in advancing our understanding of the representational format of possible verbs and the events they name. From the evidence seen so far, it would seem then that there is not much that is special about death when it comes to our verbal concepts. The verbs in (6) and (7) conform to manner/result complementarity as closely as their mundane cousins.

However, closer inspection reveals that events of death may hold a special status when it comes to the ways we can package verbal meanings together. Unlike the more mundane events of everyday life, some verbs related to events of death, like those in (13a) and (13b), appear to encode both a manner and a result in apparent violation of manner/result complementarity [5]. The verb crucify denotes both the manner in which a particular event process unfolds—that is, as a process in which an individual is affixed to and left up on a cross—along with the result of that event unfolding, which at least at first blush appears to be a result of death. The same holds for the verb disembowel, where the manner is now one in which the organs of the gastrointestinal track are removed through an incision in the belly, and the result seems again to be one of death.

• (13)

  • a. St. Andrew was crucified.

    • i. Manner: crucification
    • ii. Result: death
  • b. Dutch executioners disemboweled Balthasar Gérard.

    • i. Manner: disembowelment
    • ii. Result: death

Although such verbs appear to have a clear manner meaning component—indeed it is the manner of the event that changes between (13a) and (13b)—we can tell that the result meaning component is encoded by attempting to add a result predicate to these sentences. As shown in (14), both crucify and disembowel reject the addition of the result of death, similar to the result verbs in (11) and unlike the manner verbs in (10).

• (14)

  • a. *St. Andrew was crucified dead.
  • b. *Dutch executioners disemboweled Balthasar Gérard to death.

Verbs like crucify and disembowel are not alone. Beavers & Koontz-Garboden [5] propose that there is a class of these verbs given in (15), which I will call manner of death verbs.²

• (15) Manner of Death Verbs (to be revised): asphyxiate, behead, crucify, hang, decapitate, disembowel, drown, electrocute, eviscerate, gas, guillotine, gut, immolate, impale, poison, quarter, smother, strangle, …

The behaviour of this class of verbs is important as it reveals that certain events of death are somehow marked in human thought, licensing them to bypass the manner/result complementarity constraint that limits our more mundane verbal concepts. In this rest of this paper, I will argue that these verbs reveal how it is that a manner and a result meaning component can be bundled together in a single verbal concept and reflect on why a result of death, as opposed to other result meanings, can accompany a manner meaning. Section 2 unpacks our currently intuitive notions of manner and result meaning and §3 further investigates the manner and result meanings of manner of death verbs specifically, suggesting a revision to the class of manner of death verbs in (15). Having isolated those manner of death verbs that truly encode both manner and result meanings, §4 demonstrates that the manner and result meanings are encoded at distinct levels of meaning. Section 5 discusses the consequences of this paper for our understanding of how verbs encode meaning and why results of death are special. Section 6 concludes.

2. Manner and result in verbal concepts

Before turning our attention to a deeper investigation of manner of death verbs, we should be clear about what distinguishes manner and result meanings and how we can make such distinctions. Rappaport Hovav & Levin [7] propose that manner meanings are distinguished from result meanings by the kinds of event of change they represent: manner meanings are a type of non-scalar change and result meanings are a type of scalar change.

In this theory, a scale is formalized to be ‘a set of degrees—points or intervals indicating measurement values—on a particular dimension (e.g. height, temperature, cost, alive/dead), with an associated ordering relation’ and a scalar change is ‘a change in value of this attribute (i.e. dimension) in a particular direction along the scale, with the direction specified by the ordering relation’ [7, p.  28].³

Applying this theory, the result meaning of a result verb like clean is encoded as a unidirectional change in a value of cleanliness. When we say that Laura cleaned Billy, what we mean is that the degree of Billy's cleanliness increased as a result of the event. The manner meaning of a manner verb like scrub, however, is not encoded as some unidirectional change in a value of some dimension. Scrubbing is a set of complex actions that need not occur in any particular order and have no underlying dimension. When we say that Laura scrubbed Billy, what we mean is that a complex series of unordered actions of scrubbing occurred to Billy.

As noted above, there are a variety of linguistic diagnostics that track this difference between manner and result verbs. Here, I focus on four diagnostics: denial of result, denial of manner, implicit direct objects, and non-subcategorized direct objects using scrub as a canonical manner verb and break as a canonical result verb.

Because manner verbs encode only the specific manner of an event and not some specific result, even strongly implied results of the event can be denied by the speaker. In (16a), the potential result of scrubbing, e.g. that the bathtub becomes cleaner, can be denied. Attempting to deny the result of a result verb is markedly odd, as shown in (16b). This is taken as evidence that result verbs encode results.

• (16)

  • a. Cinderella scrubbed the bathtub, but it didn't get any cleaner.
  • b. *Harrison Ford broke his leg, but it didn't become broken.

Attempting to deny the manner of an event reveals the opposite pattern. As result verbs only encode the result of an event and not some specific manner, even strongly implied manners for the event can be denied by the speaker. In (17b), the potential manner of breaking, e.g. that of falling, can be denied. Attempting to deny the manner of a manner verb, however, is judged to be unacceptable, as shown in (17a). This is taken as evidence that manner verbs encode manner.⁴

• (17)

  • a. *Cinderella scrubbed the bathtub, but not by scrubbing.
  • b. Harrison Ford broke his leg, but not by falling.

Perhaps more interestingly, manner and result verbs differ in the way they relate to direct objects, demonstrating how native speakers tie their knowledge of verbal meanings to grammatical relations. Native speakers tacitly know that manner verbs allow them to leave the direct object implicit. In (18a), there must be some object that Cinderella is scrubbing, but the speaker is not required to tell us what it is. Result verbs, however, require native speakers to explicitly mention a direct object. In (18b), the speaker is obligated by their mental grammar to explicitly say what it was that Harrison Ford broke.

• (18)

  • a. Cinderella scrubbed (all morning).
  • b. *Harrison Ford broke (all morning).

Native speakers also tacitly know that manner verbs can also occur with so-called non-subcategorized objects, i.e. direct objects that are selected by some predicate other than the predicate of the verb. This can be seen in the interpretive difference between (19a) and (19b). In (19a), Cinderella may not be scrubbing her fingers; instead, her fingers could be in a raw state because of the scrubbing she was doing to some implicit object. This interpretation is distinct from (19b) in which Cinderella must be scrubbing her fingers (though we do not know what state her fingers end up in due to this scrubbing). Result verbs again differ in this respect and cannot occur with non-subcategorized objects. The example in (20a) cannot be interpreted as Harrison Ford's leg becoming bloodied because he broke some unspecified thing, and thus is considered unacceptable to native speakers.

• (19)

  • a. Cinderella scrubbed her fingers raw.
  • b. Cinderella scrubbed her fingers.

• (20)

  • a. *Harrison Ford broke his leg bloody.
  • b. Harrison Ford broke his leg.

These diagnostics distinguish between manner and result verbs, and in particular can be used to demonstrate when a verbal concept encodes a manner or a result meaning. In the next section, we will apply these diagnostics to manner of death verbs to uncover those verbal concepts of events of death that transcend manner/result complementarity and deserve further attention.

3. Manner of death verbs

As introduced above, manner of death verbs are verbs that appear to name both the manner in which an event of death occurs and the result of that event, namely, death.⁵ Beavers & Koontz-Garboden [5] offer the first thorough investigation of manner of death verbs and provide evidence that these verbs encode both a manner and a result component in their verbal concepts. Intuitively, these verbs express that a particular manner of action happened to an individual and that the action resulted in the death of that individual. For example, (21) expresses that drowning/hanging/electrocution/crucifixion happened to Shane and that Shane died.

• (21) Shady drowned/hanged/electrocuted/crucified Shane.

  • a. Manner: drowning/hanging/electrocution/crucifixion
  • b. Result: death

In this section, I apply the diagnostics introduced in the previous section, demonstrating that manner of death verbs encode both manner and result components in their verbal concepts.

(a). Result in manner of death verbs

Most mundane verbs appear to wear their manner or result meaning component ‘on their sleeve’, so to speak; indeed, we think of them as naming either the manner or the result of an event. The verb scrub names an event with the particular manner of scrubbing, and the verb break names an event with the particular result of breaking. Therefore, application of denial of result or denial of manner diagnostics was straightforward, and repetition of the same lexical item ensured that we were targeting a precise manner/result.

For manner of death verbs, the precise meaning of their manner component is fairly clear,⁶ but the meaning of their result component is more debatable since it is not encoded on the surface. As a first approximation, the result meaning component seems to be whatever the result meaning component of kill or die is, and indeed, applying these to a variety of manner of death verbs in (22), native speakers report the same contradictory interpretations reported for (16b).

• (22)

  • a. *The revolutionaries guillotined King Louis XVI, but he wasn't killed.
  • b. *Herennius decapitated Cicero, but he wasn't killed.
  • c. *Geoffrey of Shrewsbury quartered Dafydd ap Gruffydd, but he didn't die.
  • d.*Thích Quǎng Dúc immolated himself, but he didn't die.
  • e.*Terrorists beheaded Daniel Pearl, but he wasn't dead.

Of course, such judgments could arise from our knowledge of the world. Having one's head removed or one's body cut into four pieces seems like a sure way to end up dead. However, certain special contexts like those in (23) allow the result of death itself to be delayed or to not obtain at all.

• (23)

  • a. The executioner guillotined the man's arm off, but he lived.
  • b. Jason decapitated the hydra, but it continued to fight.
  • c. The research quartered the starfish, and it grew into four little starfish.
  • d.The asbestos monster immolated itself, but it did not die.
  • e.The farmer beheaded the chicken, but it continued to run around the yard.

The examples in (23) suggest caution when interpreting the denial of result evidence in (22). When it comes to understanding whether manner of death verbs are true counterexamples to manner/result complementarity, Beavers & Koontz-Garboden [5, p. 338] argue that the exact meaning of a manner of death verb's result component is not important, only that it has some result component because some result does obtain. For instance, the verb guillotine may not always entail death, but does entail the loss of a body part (typically the head), and therefore a result. Similarly crucify requires at least a change of location result, since crucification requires one ‘being hung upright in a particular configuration’. Indeed, this type of argument can also be made for decapitate and behead in (23) as both verbs entail the resulting loss of a head. A case can also be made that the result of quarter in (23) is the creation of four pieces. Less clear is immolate in (23) where immolation may be argued to cause no necessary change to an asbestos monster. Indeed, these arguments seem to be heavily reliant on particular world knowledge, something we may not want our analysis to rely upon.

These arguments, however, must be considered carefully, as they may threaten any coherent notion of a lexically entailed result meaning. As we push the result meaning further and further from something intuitively verb specific, all kinds of unintended but potential result meanings might be at issue. Consider some of the ‘hard-core’ manner verbs which, when pushed, certainly obtain some kind of result. The verb nibble, for instance, in (24a) entails some resulting consumption or ‘biting off’, thus resulting in an irreversible change in the apple along some scale. Similarly, the verb wipe in (24b) entails some resulting surface contact, thus there is a result of contact between the table and Mary (or the instrument that Mary uses). To my mind, these results all meet the definition of result as a scalar change given in Rappaport Hovav & Levin [7] and thus present potential problems for a clear delineation of manner and result meanings. Indeed, determining whether there is a result component and what that component actually is is a non-trival process.

• (24)

  • a. *The squirrel nibbled at the apple, but the apple remained whole.
  • b. *Mary wiped the table, but no contact with the table resulted.

In attempting to address such worries when it comes to denial of result diagnostics, Beavers & Koontz-Garboden [5, p. 337] propose a more general test aimed at detecting ‘some single, non-verb-specific inference for all result verbs’, which I will call the nothing different test. They report that application of the nothing different test does not create a contradiction with canonical mundane manner verbs, shown in (25), but does create a contradiction with canonical mundane result verbs, show in (26).⁷ A floor can be swept with no change in the floor resulting, but if a vase is broken then there has been a change to the vase.

• (25)

  • a. Tracy just swept the floor, but nothing is different about it.
  • b. Bob just yelled, but nothing is different about him.
• (26)

  • a. *Shane just broke the vase, but nothing is different about it.
  • b. *Shane just destroyed his house, but nothing is different about it.

But the nothing different test may not be as clear cut as Beavers & Koontz-Garboden [5] would like it to be. Consider again the case of canonical mundane manner verbs like nibble. If the squirrel nibbles at the apple even a little bit, then some little bit of the apple is no longer there, and it is not the case that ‘nothing is different about it’. Similarly for wipe, if Mary wipes the table, then some contact with the table must be made, and again it is not the case that ‘nothing is different about it’ as it has now had contact with something. Instead, what seems to be the case in all of these instances is that there is no relevant result such that that result is encoded into these manner verbs. That the apple is minimally consumed in a nibbling event, or that contact with the table is made in a wiping event isn't a relevant or salient result from our human perspective.

• (27)

  • a. ??The squirrel nibbled at the apple, but nothing is different about it.
  • b. ??Mary wiped the table, but nothing is different about it.

Husband [10] suggests a different approach to detecting the result component of manner of death verbs, following a more general restriction on result predicates noted by Tenny [11]: only one result is possible per (complex) event. We can see this constraint on display in (28) in which verbs that clearly result in death, namely kill and die are judged to be incompatible with the resultative to death.⁸

• (28)

  • a. *North Korea killed two civilians to death.
  • b. *Kurt Cobain died to death.

Interestingly, some manner of death verbs seem to allow the resultative to death as seen in (29), while other manner of death verbs are, as expected, unacceptable with the resultative to death as seen in (30).

• (29)

  • a. The state of Florida electrocuted Ted Bundy to death.
  • b. Joe Delaney drowned to death.
  • c. Richard Montague was strangled to death.
  • d. Michael Hutchence asphyxiated himself to death.
  • e. Napoleon was poisoned to death.
• (30)

  • a. *The revolutionaries guillotined King Louis XVI to death.
  • b. *Herennius decapitated Cicero to death.
  • c. *Geoffrey quartered Dafydd ap Gruffydd to death.
  • d. *Thích Quǎng Dúc immolated himself to death.
  • e. *Terrorists beheaded Daniel Pearl to death.

This suggests that manner of death verbs as listed in (15) in fact form two separate classes. The class of manner of death verbs in (30) appear to encode a result meaning component, thus blocking the addition of a second resultative with the same meaning like the undisputed result verbs kill and die in (28). The other class of manner of death verbs in (29), however, do not seem to encode a result meaning component. Because this latter group doesn't appear to encode a result meaning component, the addition of a resultative with the meaning to death is acceptable.

An examination of naturally occurring examples in the Google Books corpus with and without to death reveals a similar picture in a comparison of (31) and (32). Verbal uses were isolated with was, were, or been Ved both with and without to death, search numbers were summed together, and the percentage of to death uses are given in the final column. The manner of death verbs in (31) uniformly have a relative high co-occurrence with to death except for drown, while those in (32) are all relatively low in their concurrence with to death, with some exceptions for crucify, guillotine and impale. These exceptions conflict with speaker judgments, suggesting that other features of these particular cases should be explored in future research. Overall though, naturally occurring examples corroborate the use of the to death diagnostic to identify two separate classes of manner of death verbs.

	verb	was/were/been Ved	was/were/been Ved to death	%
(11)	asphyxiate	29 540	179	0.606%
	hang	1 620 000	4150	0.256%
	drown	2 750 000	710	0.026%
	electrocute	73 400	879	1.198%
	gas	87 900	4266	4.853%
	poison	665 300	5443	0.818%
	smother	151 800	5850	3.854%
	strangle	203 900	15 020	7.366%
(12)	behead	679 500	9	0.001%
	crucify	1 165 200	772	0.066%
	decapitate	110 100	2	0.002%
	disembowel	29 920	0	0.000%
	eviscerate	23 340	1	0.004%
	guillotine	64 700	83	0.128%
	gut	125 800	4	0.003%
	immolate	26 780	0	0.000%
	impale	64 500	64	0.099%
	quarter	382 100	2	0.001%

Why might some supposed manner of death verbs occur with to death resultatives? If we examine the set of manner of death verbs that occur in (29) and (31), they are the kinds of incidents that are, in fact, known to be survivable. An individual can be electrocuted, for instance, by being struck by lightning, and yet survive to tell the tale. There are also cases of individuals being hanged, only to be cut down and revived before dying. Therefore, the events that these verbs name are not true manners of death, but instead are simply events with a manner that can, at times, lead to death. As such, these verbs are more like the manner verbs shot and stab in (6).

The to death diagnostic acts as a strong test for true manner of death verbs. To the extent that native speakers find the combination of a manner of death verb unacceptable with to death, there is good reason to claim that that verb encodes a result. To that end, a revised list of manner of death verbs is given in (33).⁹

• (33) Manner of Death Verbs (revised): behead, crucify, decapitate, disembowel, eviscerate, guillotine, gut, immolate, impale, quarter, …

Turning to other diagnostics, native speaker intuitions about the grammar of manner of death verbs are also revealing. Like more mundane result verbs, manner of death verbs do not allow implicit direct objects. As the examples in (34) demonstrate, the individual who is being guillotined or quartered or beheaded must be overtly mentioned. This is further evidence that manner of death verbs encode a result as part of their verbal concept.

• (34)

  • a. *The revolutionaries guillotined.
  • b. *Herennius decapitated.
  • c. *Geoffrey of Shrewsbury quartered.
  • d. *Thích Quǎng Dúc immolated.
  • e. *Terrorists beheaded.

The inability to occur with non-subcategorized objects also favours the encoding of a result meaning component in manner of death verbs. The examples in (35) demonstrate that the object of manner of death verbs cannot be selected by a predicate that is not the verb. Thus, while is it sensible to think that decapitation is a good way to silence one's enemies or that self-immolation is the ultimate of compassionate acts to resist religious intolerance, a speaker can not express this using (35b) or (35d).

• (35)

  • a. *The revolutionaries guillotined King Louis XVI off the throne.
  • b. *Herennius decapitated Cicero silent.
  • c. *Geoffrey of Shrewsbury quartered Dafydd ap Gruffydd into exile.
  • d. *Thích Quǎng Dúc immolated himself compassionate.
  • e. *Terrorists beheaded Daniel Pearl kidnapped.

(b). Manner in manner of death verbs

Having established more clearly the result component found in the more restricted set of manner of death verbs in (33), a second challenge arises concerning the status of the manner component of these verbal concepts. In a careful exploration of the purported manner of death verb drown, Rappaport Hovav [16] argues that some manner of death verbs may, in fact, simply be result verbs that lack a manner. For drown, she argues that it encodes only a result, i.e. that someone/something becomes submerged, often in water. Although the verb drown was already removed from the list of manner of death verbs in (33), other verbs on this list may also be threatened. For instance, as noted above, the verbs decapitate and behead entail the loss of a head, and these entailed meanings might, in fact, simply be results as they appear to encode a scalar change from having to not having a head. Similarly, quarter may too only encode a scalar result that something/someone is separated into four pieces. Stronger evidence therefore is needed to secure the manner component of manner of death verbs.

An initial diagnostic for the presence of a manner meaning in a verbal concept involves denial of the manner meaning. As noted above, the manner component of manner of death verbs is often fairly clear, allowing the denial of manner test to be more straightforwardly applied and interpreted. As manner of death verbs differ from one another specifically in the manner in which the event they name unfolds, it is unsurprising that attempts to deny their named manner, shown in (36), results in contradiction. For those manner of death verbs where the manner meaning can be spelled out in more detail, denial of the subprocesses of the manner meaning are also contradictory, as in (37). These denial diagnostics are initial evidence that manner of death verbs encode a manner component as part of their verbal concept.

• (36)

  • a. *Geoffrey of Shrewsbury quartered Dafydd ap Gruffydd, but not by quartering him.
  • b. *Terrorists beheaded Daniel Pearl, but not by beheading him.
• (37)

  • a. *The revolutionaries guillotined King Louis XVI, but not by releasing the rope and letting the blade drop.
  • b. *Herennius decapitated Cicero, but not by striking several times with a sword at his neck.
  • c. *Thích Quǎng Dúc immolated himself, but not by setting himself on fire.

Beavers & Koontz-Garboden [5] suggest other diagnostics to identify and isolate the manner component of a verbal concept. One diagnostic involves the relationship between a verb and its subject. Result verbs place few requirements on their subjects, allow human agents like John, instruments like a hammer, or natural forces like an earthquake to be subjects, as shown in (38). Manner verbs, however, place more requirements on their subjects, so human agents like John can appear as subject, but not instruments like a stiff brush or natural forces like an earthquake, as shown in (39).

• (38)

  • a. John broke/shattered the vase (with a hammer).
  • b. The hammer broke/shattered the vase.
  • c. The earthquake broke/shattered the vase.
• (39)

  • a. John scrubbed/wiped the floor (with a stiff brush).
  • b. ??The stiff brush scrubbed/wiped the floor.
  • c. ??The earthquake scrubbed/wiped the floor.

Importantly, it is not that instruments cannot appear with manner verbs. A comparison of (39a) and (39b) demonstrates that this restriction is more specifically related to instruments being subjects. Additionally, as Beavers & Koontz-Garboden [5] note, natural forces are not omitted from manner verbs simply because they could not plausibly function as a direct cause for scrubbing/wipping. If a wet towel were on the floor at the time of an earthquake, it could be moved around by the earthquake, thereby scrubbing/wiping the floor. However, even with this scenario in mind, (39c) is judged to be odd. Instead, manner verbs like scrub and wipe appear to encode a meaning that involves a specific method of applying an instrument to a patient.

Turning to manner of death verbs, these also appear to place strict requirements on their subjects, allowing agents like revolutionaries, but not instruments like a rusty blade or natural forces like a heavy wind. Again in these cases, a comparison of (40a) and (40b) shows that the instrument can appear with the verb, just not as its subject, and for natural forces in (40c), a direct cause is plausible, but it fails to make these cases acceptable.

• (40)

  • a. The revolutionaries guillotined King Louis XVI (with a rusty blade).
  • b. ??The rusty blade guillotined King Louis XVI.
  • c. ??A heavy wind guillotined King Louis XVI (by loosening the blade).

A second diagnostic they deploy involves distinguishing the temporal duration of the manner component from the temporal duration of the result component. When an event requires a complex process to unfold, the complexity of this set of processes requires a non-trivial interval of time to elapse. Because of this requirement, complex meanings are temporally durative.

Temporal durativity can be diagnosed with a took time test [17]. In particular, a durative ‘during 5 min’ interpretation of take 5 min can occur with a complex event. Simple events only allow for a punctual ‘after 5 min’ interpretation to occur. For manner verbs, take 5 min can only be interpreted as ‘after 5 min’ for a simple manner action like blink in (41a), but is ambiguous between an ‘after 5 min’ or ‘during 5 min’ interpretation for a complex manner action like run in (41b). For result verbs, Beavers [15] furthermore demonstrates that take 5 min can only be interpreted as ‘after 5 min’ for a simple two-point scalar change result like break in (42a), but is ambiguous between an ‘after 5 min’ or ‘during 5 min’ interpretation for a complex multi-point scalar change result like clean in (42b).¹⁰ Therefore, to use this test to isolate a manner meaning for a possible verbal concept encoding both manner and result meanings, we need to know both whether the manner of the verbal concept is simple or complex and whether the result of the verbal concept is simple or complex.

• (41)

  • a. It took John 5 min to blink. (‘after’ only)
  • b. It took John 5 min to run. (‘after’ or ‘during’)
• (42)

  • a. It took John 5 min to break the vase. (‘after’ only)
  • b. It took John 5 min to clean the vase. (‘after’ or ‘during’)

For manner of death verbs, the result component ultimately represents a simple two-point scale: not dead or dead.¹¹ Because of this, we would expect a take time test to only permit an ‘after’ interpretation if these verbs merely encoded this simple result. However, some manner of death verbs also permit a ‘during’ interpretation, and this interpretation must be linked to a complex manner.¹² Beavers & Koontz-Garboden [5] observe that the verb crucify in (43a) can have a durative interpretation in addition to an ‘after’ interpretation. On its durative interpretation, the complex non-scalar event process encoded in the manner meaning of crucify is highlighted: an individual is tied/nailed down, hoisted upright and then monitored until death. Similarly, a durative interpretation of disembowel highlights a distinct complex non-scalar event process, requiring a horizontal incision to the belly, removal of the gastrointestinal track though the incision and monitoring until death.

• (43)

  • a. It took Aegeas' men three days to crucify St. Andrew …
    after three days: because that is how long it took for St. Andrew to die.
    during three days: because that is how long it took to tie St. Andrew down, hoist him up and wait for his death.
  • b. It took executioners several hours to disembowel Gérad …
    after several hours: because that is how long it took Gérad to die.
    during several hours: because that is how long it took to make an incision in Gérad's belly, remove his intestines and wait for him to expire.

Importantly for both crucify and disembowel, the complex manner highlighted by the durative interpretation is idiosyncratic to the particular verb, suggesting that these complex manners must be specifically encoded in these particular verbal concepts.

Taken together, the diagnostics reviewed above along with others developed and applied in Beavers & Koontz-Garboden [5] strongly support the argument that manner of death verbs are counterexamples to manner/result complementarity. Moreover, the status of manner of death verbs as counterexamples to what is otherwise a well-supported constraint on verbal meaning suggests that manner of death events are somehow specially marked in human thought. In the next section, however, I will outline a key distinction between the manner and result components of manner of death verbs.

4. Assertion, presupposition and manner of death verbs

In defiance of manner/result complementarity, manner of death verbs encode both a manner and a result meaning component. However, the idea of encoding that we have been using up to this point has been somewhat unspecified and has treated meaning as a single unified phenomenon. However, research in semantics and pragmatics has found that the variety of meanings that are composed together in a single sentence are often not of a single type. Instead, sentence meaning can be split apart on several different levels of meaning. One question to ask is whether lexical meanings (i.e. the meaning components of verbs) can be similarly split apart into multiple types, each possibly encoded on a different level. Here, I will focus on two types of meaning: meaning that is asserted, and meaning that is presupposed.

Asserted meaning is the meaning of a sentence that is directly expressed. It is often what is taken to be the at-issue or new content of a sentence that is foregrounded by the speaker. Presupposed meaning, on the other hand, is the meaning of a sentence that is more indirectly expressed. It is often taken to be meaning that is not-at-issue and may be backgrounded as part of what speakers take to be uncontroversial and not contested, i.e. the common ground between speakers.

Consider the case of a sentence with the aspectual verb stop given in (44), which bears two meanings given in (44a) and (44b). These two meanings have different status. The meaning in (44a) is the meaning that is asserted by the speaker. This meaning is what is said to be at-issue, bringing in new information to the unfolding discourse. The meaning in (44b), however, is something that the speaker is taking for granted, that is, the speaker presupposes that smoking used to occur.

• (44) John stopped smoking

  • a. Assertion: John does not smoke now.
  • b. Presupposition: John used to smoke.

As with the distinction between manner and result meanings discussed above, perhaps the clearest way to distinguish assertion from presupposition is by way of their behaviour with respect to certain linguistic diagnostics. The classic diagnostics for identifying asserted versus presupposed meanings are their behaviour with questions and negation: assertions can be directly questioned and negated; presuppositions cannot [18].

When a yes/no question is asked of (44) as in (45), native speakers take the question being asked to only be about (44a), whether or not John smokes now. The question cannot be asking about (44b), whether or not John used to smoke. Intuitively, it would be very misleading for a speaker to ask (45) if, in fact, John has never smoked. Instead, the proposition that John used to smoke in (44b) is taken to be common ground. Thus, a response to the question that attempts to deny that any smoking ever occurred seems like an unacceptable continuation of the discourse.

• (45) Speaker A: Did John stop smoking?

  • a. Speaker B: No, John still smokes. (Assertion)
  • b. Speaker B: *No, John has never smoked! (Presupposition)

Similarly, negation of (44) as in (46) cannot negate (44b), that John used to smoke, but only (44a), that John doesn't smoke now. As with questions, a continuation of the discourse that attempts to argue that no smoking ever took place is unacceptable.¹³

• (46) Speaker A: John did not stop smoking.

  • a. Speaker B: Wrong. He did stop smoking. (Assertion)
  • b. Speaker B: *Wrong. He has never smoked! (Presupposition)

Thus, presuppositions have a special status when it comes to sentence-level meaning operators like questions and negation. Because they cannot be the targets of these sentence-level meaning operators, presupposed meanings are said to project outside of the scope of questions and negation. Assertions lack this property, and are thus targeted by questions and negation.

Given this difference in the status of asserted compared to presupposed meaning and the possibility that different meaning components could be encoded on different levels of meaning, it is important to establish what types of meaning the manner and result meaning components of manner of death verbs are.

Beginning with the manner component of manner of death verbs, the discourses given in (47) are evidence that the manner component of manner of death verbs does not project out of questions. When Speaker A asks a yes/no question, Speaker A is asking about the manner in which the individual died. This can be seen in that the responses by Speaker B, which may affirm the manner or deny it and propose a different manner, are acceptable continuations for discourse.

• (47)

  • a. Speaker A: Was King Louis XVI guillotined?
    Speaker B: Yes, he was killed by a guillotine.
  • b. Speaker A: Was Cicero decapitated?
    Speaker B: No, he was stabbed.
  • c. Speaker A: Was Dafydd ap Gruffydd quartered?
    Speaker B: Yes, he was cut up into four pieces.
  • d. Speaker A: Did Thích Quǎng Dúc immolate himself?
    Speaker B: Yes, he set himself on fire.
  • e. Speaker A: Did terrorists behead Daniel Pearl?
    Speaker B: No, he was shot in the back of the head.

The manner component of manner of death verbs also does not project out of negation, as seen in the discourses given in (48). When a speaker negates these sentences, he is negating the particular manner being asserted by the verb. Proposing a different manner in light of this negation is, thus, acceptable when continuing a discourse.

• (48)

  • a. King Louis XVI was not guillotined. He was strangled to death!
  • b. Cicero was not decapitated. He was stabbed!
  • c. Dafydd ap Gruffydd wasn't quartered. He was drawn and hung.
  • d. Thích Quǎng Dúc didn't immolate himself. He died in his sleep.
  • e. Terrorists didn't beheaded Daniel Pearl. He was killed by firing squad.

Turning to the result component of manner of death verbs, the discourses given in (49) are evidence that the result component of manner of death verbs projects out of questions. By asking a yes/no question, Speaker A is not questioning whether the individual died. This can be seen in that the responses by Speaker B, which affirm or deny that the individual's death are unacceptable continuations for discourse.

• (49)

  • a. Speaker A: Was King Louis XVI guillotined?
    Speaker B: *No, he did not die.
  • b. Speaker A: Was Cicero decapitated?
    Speaker B: *No, he did not die.
  • c. Speaker A: Was Dafydd ap Gruffydd quartered?
    Speaker B: *Yes, he was killed.
  • d. Speaker A: Did Thích Quǎng Dúc immolate himself?
    Speaker B: *No, he did not die.
  • e. Speaker A: Did terrorists behead Daniel Pearl?
    Speaker B: *Yes, he died.

The result component of manner of death verbs also projects out of negation, as seen in the discourses given in (50). When a speaker negates these sentences, he cannot be directly negating the individual's death. Thus, expressing that the individual continued to live immediately after negating the manner of death verb sentence is unacceptable when continuing a discourse.

• (50)

  • a. King Louis XVI wasn't guillotined. *He remained King of France for years.
  • b. Cicero wasn't decapitated. *He led the revolution against Caesar.
  • c. Dafydd ap Gruffydd wasn't quartered. *He escaped to France.
  • d. Thích Quǎng Dúc didn't immolate himself. *He's alive in Vietnam.
  • e. Terrorists didn't behead Daniel Pearl. *He's an active journalist in the Middle East.

Evidence from questions and negation point to differences in the manner and result meanings encoded in manner of death verbs. The manner component is an assertion; the result component is a presupposition. Importantly, both assertions and presuppositions are lexical entailments, i.e. meanings that must be encoded in a verbal concept [19]. As such, manner of death verbs remain counterexamples for the original specification of manner/result complementarity given in (12). The distinction between the levels of meaning of each of these components, however, suggests an important difference in the way they are encoded within a single verbal concept.

5. Manner/result complementarity and the special status of death

The previous two sections have demonstrated that manner of death verbs encode both manner and result meaning components and that these components reside at different levels of meaning: the manner component is asserted, and the result component is presupposed. Given this analysis, two issues remain to be addressed. The first issue concerns the status of manner/result complementarity as a constraint on verbal meanings. As manner of death verbs are counterexamples to manner/result complementarity as presented in (12), how should we think about possible constraints on verbal meanings? The second issue concerns the seemingly special status of death. Why is it that results of death can be accommodated into complex verbal concepts when other types of results fail to do so? I'll address each of these in turn.

(a). Complementarity of assertions

Considering the implications of manner of death verbs as counterexamples to manner/result complementarity, one tempting avenue is to abandon any hard grammatical constraint on verbal concepts. However, the fact that the manner and result components of manner of death verbs reside on different levels of meaning suggests an alternative pathway. Complementarity may be a hard constraint after all, but perhaps it is a constraint only on certain levels of meanings.

Widening our view somewhat, manner of death verbs are not the only verbs that have some kind of complex verbal concept. Consider the classic case of a verb like regret in (51) when compared with dream in (52). In addition to their more overt meanings (disappointment and dreaming, respectively), both verbs take propositions like he smoked as complements, but the status they give to their propositions is distinct. The verb regret requires its proposition to be a fact, something true of the real world, while dream has no such requirement. Saying (51) assumed the proposition he smoked to be true. Therefore, we say that regret is a factive verb.

• (51) Sam regrets that he smoked.

  • a. Assertion: Sam is disappointed that he smoked.
  • b. Presupposition: Sam has smoked.
• (52) Sam dreamed that he smoked.

  • a. Assertion: Sam had a dream that he smoked.
  • b. Presupposition: Sam has smoked.

Furthermore, we can demonstrate that the disappointment meaning and factive meaning carried by regret reside on different levels of meaning. If a speaker asks a yes/no question of (51), as in (53), the speaker is asking about Sam's disappointment, not about the underlying fact of his smoking. Similarly, if the speaker negates (51), as in (54), they are negating Sam's disappointment, not the underlying fact of his smoking. Also, attempts to carry on the discourse by directly addressing the factive meaning are considered unacceptable to speakers. Therefore, the factive meaning of regret is encoded as a presupposition because it projects out of questions and negation.

• (53) Does Sam regret having smoked?

  • a. No, he isn't disappointed that he smoked. (Assertion)
  • b. *No, he didn't smoke. (Presupposition)
• (54) Sam doesn't regret having smoked.

  • a. He enjoyed smoking a lot. (Assertion)
  • b. *He didn't smoke at all! (Presupposition)

A different case comes from the treatment of the verb acquiesce in (55), which also has at least two components to its meaning: an acceptance meaning and a reluctance meaning. This suggests that acquiesce also encodes a complex underlying verbal concept.

• (55) Bill acquiesced to our demands.

  • a. Assertion: Bill accepted our demands.
  • b. Presupposition: Bill was reluctant to accept our demands.

As with regret, only one of these two meaning components is asserted. A speaker questioning (55) as in (56) can only be directly asking about Bill's acceptance of the demands, not about Bill's willingness to accept the demands. The same pattern is revealed by a speaker negating (55) as in (57), with negation only able to target Bill's acceptance of the demands, not his reluctance to accept them.

• (56) Did Bill acquiesce to our demands?

  • a. No, Bill did not accept our demands. (Assertion)
  • b. *No, it was easy for Bill to accept our demands. (Presupposition)
• (57) Bill did not acquiesce to our demands.

  • a. He did not accept a single one. (Assertion)
  • b. *He found it easy to accept our demands. (Presupposition)

Looking beyond verbs for the moment, the much discussed noun bachelor is known to have many components to its meaning. To appropriately apply bachelor to an individual, that individual needs to be someone who is unmarried, human, male, of marrying age, not previously divorced, and not restricted from marriage. When we probe each of these meanings, we find that speakers presuppose most of these attributes. The only component that is asserted is the unmarried meaning component. As seen in dialogues with questions in (58) and negation in (59), all other components (human, male, of-marrying-age, not-divorced, and possibility-of-marriage) are found to be presupposed (P. Jacobson 2010, personal communication).

• (58) Is Jessie a bachelor?

  • a. No, he's married to Clare.
  • b. *No, he's a humpback whale!
  • c. *No, Jessie's a girl!
  • d. *No, he's a 5 year old!
  • e. *No, he's a divorcée!
  • f. *No, he's a catholic priest!

• (59) Jessie's not a bachelor

  • a. He's married to Clare.
  • b. *He's a humpback whale!
  • c. *Jessie's a girl!
  • d. *He's a 5 year old!
  • e. *He's a divorcée!
  • f. *He's a catholic priest!

The generalization behind these examples and those concerning manner of death verbs is that complex lexical concepts assert only one component of their meaning, with other meaning components forming some kind of presupposition. In the light of these observations, a revision to the initial hypothesis as it pertains to manner/result complementarity in verbal meaning is proposed in (60), which restricts complementarity to the assertive level of meaning.

• (60) Manner/result complementarity (revised): Manner and result meaning components are in complementary distribution with respect to the asserted level of meaning: a verb can assert only one.

This revised version of manner/result complementarity raises several questions that I would like to briefly explore. First, given (60), we would expect to find classes of verbs that assert their result and presuppose their manner, though at present no cases are known to me.¹⁴ While this possibly suggests a further refinement of manner/result complementarity, another potential explanation comes from event structure differences between manners and results. Verbs asserting a result and presupposing a manner may be unattested because of the way our grammar encodes different event structure components: manner verbs are modifiers of event structure, whereas result verbs are arguments of event structure [20]. Modifiers and arguments have a different status in the grammar more generally. By way of explication, modifiers like the adverb yesterday in (61) are always optional elements. They are not inferred when omitted and are not covert, so speakers do not infer yesterday in (61b) or (61d) where yesterday is absent. Arguments like something in (61), however, can in certain cases be inferred when omitted and may be covert. So speakers can infer that something was eaten in (61c) and (61d) even though something is absent.

• (61)

  • a. John ate something yesterday.
  • b. John ate something.
  • c. John ate yesterday.
  • d. John ate.

Because the presence of a modifier is always optional and, thus, unpredictable, it may take on a more central role in meaning. This difference may interact with the process of verb learning, blocking certain meaning combinations from easily arising. While not currently worked out in detail, consider the following sketch of a theory of this interaction. When learning the meanings of a verb with a modifier and a result meaning, the modifier meaning of the verb must be asserted because modifiers are grammatically unpredictable. With its asserted meaning filled, the result meaning of the verb must be presupposed because of manner/result complementarity, and can be presupposed given its status as an argument.

Another potential issue concerns the limits, if any, on the number of presuppositions a word can encode. Given attested examples like bachelor, it seems that the number of presuppositions a lexical item can have may be potentially large. Indeed, even if the attested upper bound were discovered to be five, as found with bachelor, it would seem that this limit would to be more of an issue concerning the limits of human learning, memory and attention than a hard constraint coming from our mental grammars. Such a division of labour between our memory and attentional systems and our mental grammars is reminiscent of other arguments for the competence/performance distinction [21]. While our grammatical competence specifies no constraint on the content of a word's presupposition, permitting an unlimited number of meaning components at that level of meaning, the number that is attested in natural language is the result of the limits of the memory and attentional capacity of the performance systems that we use to acquire, comprehend and produce language.

Finally, there is a wider question as to why languages seem to have very few verbal concepts that encode both a manner and a result component. One possibility, again drawing on preferences related to human memory and attention, is that native speakers may tacitly prefer to minimize the number of meaning components encoded in a verb given the morphology of their language.¹⁵ While English is relatively poor morphologically, we may see evidence of this in Dutch where some manner of death verbs require a verbal prefix that combines with a verb stem to create a result interpretation, as given in (62).¹⁶

• (62)

  • a. ont-hoofden ‘decapitate/behead’ hoofden ‘head’
  • b. ver-drinken ‘drown’ drinken ‘drink’
  • c. ver-stikken ‘asphyxiate’ stikken ‘choke’

Arsenijevic̀ [23, p. 5] makes similar observations about verbal prefixes in Serbo-Croatian. He proposes that this prefix tells speakers that ‘the strangling event took place in its full extent, with its characteristic phase transition of Marija switching from being alive to being dead.’

• (63)

  • a. Jovan je gušio Mariju.
      Jovan aux strangle Marija
      ‘Jovan was strangling Marija.’
  • b. Jovan je u-gušio Mariju.
      Jovan aux in-strangle Marija
      ‘Jovan strangled Marija (to death!).’

As these morphologically rich languages can encode results through a verbal affix, constructing a verbal concept expressing both a manner and a result without using a verbal prefix may be dispreferred. Of course, the addition of a result meaning by verbal prefix may change the status of the result meaning itself from one that is presupposed by the speaker to one that is now asserted, and understanding the extent of such differences requires further cross-linguistic evidence and exploration.

(b). The special status of death

Why is it that, of all the results that we humans witness and encode in our languages, events that result in death appear to stand out? Indeed, aside from the class of manner of death verbs discussed here, other examples of violations of manner/result complementarity are in rather short supply.¹⁷ The potentially unique and certainly exceptional status of the result of death in manner of death verbs raises some important questions: what is required for an additional meaning component to be accommodated in a verbal concept, and why are results of death able to meet these requirements?

At least two requirements appear to be at play. First, the meaning component must be important and prominent to human interest. Second, the component must satisfy the constraints of our mental grammars. On the first, additional meaning components clearly need to be salient to human thought to be accommodated into a more complex verbal concept. If a meaning component is not salient, then it is highly unlikely to become encoded in a verbal concept, simply again because of the limits on human memory and attention in place during word learning. Salience is very vague requirement, but consider another case where a similar notion appears to be at play, the rather muddled realm of implicit arguments of verbs. Take the verb steal, for example. Native speakers know that every event of stealing requires a thief, a victim, and some loot. They also know that events of stealing must take place at a particular time. Speakers are not required to explicitly mention all of these arguments. Given (64b) to (64d), we can see that speakers can leave the victim, the time or both implicit. However, when left unmentioned, the status of some arguments seems more salient compared to others. In (64d) the victim and the time of the stealing event remain implicit, and yet it is intuitively clear that the implicit victim of the event is far more salient in our minds than the implicit time of the event, even though both arguments are entailed by the event itself.

• (64)

  • a. Valerio Viccei stole £60 million from Knightsbridge Safe Deposit Centre on 12 July 1987.
  • b. Valerio Viccei stole £60 million on 12 July 1987.
  • c. Valerio Viccei stole £60 million from Knightsbridge Safe Deposit Centre.
  • d. Valerio Viccei stole £60 million.

Williams [28, p. 85] proposes that arguments that are either required overtly or implied by a verb make up a mental representation he calls a sketch, similar to earlier ideas found in Fillmore [29,30], Langacker [31] and many others. In his terms, a sketch ‘provides a psychological perspective on things that satisfy the predicate, perhaps engaged by default when one thinks of a thing as a satisfier of the predicate’. Thus the verbal concept of steal engages the notion of a victim by default as part of a stealing event, but not the notion of a time that resides elsewhere.

In a similar way, the verbal concepts for manner of death verbs also engage with a result of death by default and seem to be part of a manner of death verb's sketch.¹⁸ Native speakers know that every event named by a manner of death verb results in death by default, and this result of death is certainly salient enough in our minds when we think of events of beheading, guillotining or immolation to, for instance, block to death resultative, as seen in (30). Importantly, it is not a grammatical requirement that manner of death verbs exist in natural language. Instead, it is this functionalist element of prominence to human interest that highlights the result of death component and enables it to be accommodated into a manner of death verbal concept.

Where grammatical requirements play an important role, however, is in the treatment of how these additional meaning components are incorporated into a complex verbal concept. This is the second requirement that any additional meaning component must also be able to satisfy our grammatical restrictions on lexical encoding. Our mental grammar gives us the resources to format a lexicon, but as with any format, certain constraints must be met. For manner of death verbs to overtly name, and thus assert, the manner of an event, manner/result complementarity as revised in (60) required us to either background any other prominent meaning component as a presupposition or abandon the attempt to incorporate it. We can see both of these routes being taken by verbs that name events that often result in death. As noted in §3, events named by verbs like electrocute, drown, strangle or even hang are, in fact, potentially survivable, thus even for those cases where death does occur, we do not accommodate a presupposed result of death into their verbal concepts. But for events named by true manner of death verbs given in (33), a result of death appears unavoidable.¹⁹ To meet manner/result complementarity, the result of death component is accommodated as a presupposition in these verbal concepts. Adding the result component in this way satisfies the manner/result complementarity constraint.

Thus it appears that both the formal constraints of our mental grammars and the functionalist pressures related to broader human interests and concerns conspire to give us the lexicons that we have. Our verbs are required to satisfy certain constraints on their formats that, by hypothesis, are dictated by manner/result complementarity in (60), and within those constraints, we encode the meanings that capture our attention and are prominent to our interests. Results of death certainly rise to the necessary level of importance in manner of death verbs, suggesting that results of death hold a special status in human thought.

6. Conclusion

A close inspection of the verbal concepts that name events surrounding death has revealed some of the exceptional ways that we think about events of death, providing a unique perspective on our mental representations of thanatological phenomena. The discovery of manner of death verbs as a special verbal class that violates the long-held manner/result complementarity constraint on verbal meaning suggests that results of death hold a special status in human thought and interest. In this paper, we have examined the manner and result components of manner of death verbs, finding that they are, indeed, encoded by these verbal concepts. Given that these verbs appear to name manners, the fact that results of death are also encoded was surprising and can be seen as further evidence that human cognition treats events of death as prominent and highly salient. Having a highly salient result but also needing to name a manner, manner of death verbs bypass the manner/result complementarity enforced by our mental grammar and constraints on learnability, encoding these meaning components on different levels of meaning. The asserted meaning of manner of death verbs encodes their named manner, while presupposing the prominent result of death meaning component. Results of death, it seems, are of such an exceptional status to our minds that we presuppose them when talking about the manners of events that result in death.

Data accessibility

This article has no additional data.

Competing interests

I declare I have no competing interests.

Funding

I received no funding for this study.