Event structure: a force dynamics / absolute construal account

Henryk Kardela, UMCS

In Ronald Langackers model of cognitive grammar, an event can be viewed from two
perspectives: from the energetic chain perspective (or force dynamics perspective) and from
the absolute (or zero construal) standpoint. Under the force dynamics construal, an event is
seen as involving the flow of energy which is transmitted from one participant to another
(Langacker 1991b: 283):
(1)

Head Tail

The initial element of the chain, which is called the head of the chain, imparts energy to the
second element in the chain, the second element relays this energy to the third until the last
element of the chain, the tail, is reached. The energetic chain, presented in (2), gives rise to
the so-called canonical event model, which underlies the prototypical transitive construction
and which, as Langacker (1991b: 285; 286) puts it, represents the normal observation of a
prototypical action:
(2)

AG PAT
setting

The canonical event model includes elements which provide specifications for semantic roles
such as agent, patient, instrument, experiencer, etc. In contradistinction to traditional
approaches, however, these roles, called by Langacker role archetypes, are treated by
cognitive grammar not as linguistic roles but rather as pre-linguistic conceptualisations. The
archetypal agent is thus a person who volitionally initiates physical activity as a result of
physical the transfer of energy takes place to an external object. The archetypal patient
absorbs the energy and undergoes a change of state (represented as a wavy arrow in (2)).
The instrument (absent in (2)) is a physical object which is manipulated by the agent and
serves as the intermediary in the transmission of energy. The experiencer (not present) is a
person who is engaged in mental activity, and the mover (or theme) is an entity which changes
its location.
Note that (2) represents the prototypical transitive construction: the participant, who is
typically an agent, initiates the energy which is transmitted onto the patient. However, there
exist constructions which depart from their transitive prototype; they form a hierarchical
organization in which each instance of such departure represents a less transitive type. The
departures from the prototype form a cline, represented by the transitivity hierarchy
(Maldonado 1992: 63):

(3) the transitivity hierarchy

transitive > reflexive > oblique intransitive > middle > intransitive absolute

The hierarchy in (3) can be exemplified by the following examples from Polish (see also
Maldonados 1992: 63 analysis of Spanish data):

(4) Marysia zapamitaa sowa Piotra. (transitive)

Mary remembered Peters words.
(5) (Ja) zapamitaem siebie jako maego chopca skorego do bjki. (reflexive)
I remembered himself as a boy ready to fight.
(6) Piotr dugo zastanawia si nad problemem przeludnienia. (oblique intransitive)
Peter reflected on the problem of overpopulation for a long time.
(7) (Ja) pamitaem o jego sowach. (middle)
I remembered (of) his words.
(8) Dugo mylaem zanim napisaem pierwsze zdanie. (absolute)
I reflected a lot before writing the first sentence.

According to (3), the most prototypical transitive construction is (4), since the energy flows
from the agent, which is coded as a subject Marysia, to the patient, coded as the direct object
Piotra sowa Peters words. A less prototypical construction is (5), in which one participant
is coded by both the subject, ja I, and the reflexive siebie myself. The least prototypical is
(7), in which only one participant appears. The absolute construal construction in (8) ends the
hierarchy; it codes the scene in the most neutral terms and constitutes a point of departure
for the alternative way of viewing an event, namely, through the prism of the absolute
construal of events. Thus consider now the following examples:

(9) a. Kiedy leje (deszcz) dzieci bawi si w domu.

When it pours with rain, the children play indoors.
b. *Kiedy si rozleje (deszcz) dzieci bd bawi si w domu.

(10) Mleko si rozlao po kuchence.

The milk spilled all over the cooker.
(9a) exemplifies the absolute construal of events: the scene of raining is construed as a normal
weather phenomenon, with no reference as to whether the rainfall was expected or not; (9a)
thus can be seen as providing the most neutral description of the event of raining. The
ungrammaticality of (9b) stems precisely from the fact that, contrary to expectations, natural
static forces of nature such as rainfall are portrayed as gathering dynamicity. In contrast, it
is perfectly natural to view (10) as involving an element of dynamicity, since the event of the
milk spilling creates a sense of unexpectedness. What signals the dynamicity here is the
reflexive si. Consider:

(11) *a. Nagle padao-Imperf.

Suddenly it rained.
b. Nagle si rozpadao-Perf.
Suddenly it (started to) rain(d).

(11a) is an equivalent of (9a), whose ungrammaticality stems from the fact that the absolute
construal is unnaturally dynamicised by the presence of the adverb nagle suddenly. In
contrast, the adverb nagle is perfectly legitimate in sentence (11b) as it portrays the
suddenness of the rainfall.
Interestingly, the same contrast: lack of energy vs. suddenness/dynamicity of event as
signalled by a reflexive is displayed by Spanish data. Consider the following examples from
Maldonaldo (1992: 537):

(12) a. No quiero desayunar ahora, prefiero esperar a que Valeria se despierte.

I dont want to have breakfast now. I would rather wait until Valeria wakes up.
b. Me (*) despert a media noche y ya no me (*) pude dormir.
I woke up in the middle of the night and I couldnt go back to sleep.

(13) No (*me) pude dormir durante una semana.

I couldnt sleep for a week.

Notice that the lack of the reflexive me in (12b) leads to ungrammaticality as it is

precisely the presence of the reflexive me here that designates (or profiles) the change of state:
from sleep to waking up. In contrast, in (13) the presence of the reflexive is illegitimate
because sleeping is a static event, stretching, in this case, over a week. What this means is that
the Spanish reflexive particle se, just as the Polish particle si, reflects the dynamic character
of the event construal. Such an increasing degree of dynamicity can be seen in (10), (11b) and
(12b), which represent middle voice constructions, constructions which, according to
Maldonado, provide a link between absolute and transitive constructions. The question now
arises as to how to account for the dynamicity in these examples. In order to answer this
question we have to address the problem of the so-called upstream flow of energy and the
A/D functional layering.
First note that whereas in the energetic chain in (1) the energy flows down-stream,
from the agent to the patient, the energy in the absolute construal flows up-stream. This is
so because the causer which dynamicises the event comes from the outside. One way of
accounting for this process is to assume that the causation involved in the process is
determined by the co-called A(utonomy)/D(ependence) functional layering in the sense of
Langacker (1991b). Because A/D layering involves what Langacker calls role archetypes, a
brief digression on the notion of role archetype is in order here.
In Langacker (1991a, 1991b) a distinction is drawn between thematic relations and
thematic relationships. Whereas the former term is used to cover any semantic role such as
agent, patient, instrument, etc., the latter refers to conceptually autonomous event components
and thematic roles which instantiate these relationships, excluding the roles of agent or
instrument. The thematic roles which enter into thematic relationships can be represented as
follows (Langacker 1991b: 288):

(14)
ZERO MOVER PATIENT EXPERIENCER THEME

Given the distinction between thematic relations on the one hand and thematic
relationships (and thematic roles) on the other, an event can, as already mentioned, be looked
upon from two perspectives: the force-dynamic construal, represented by the flow-of-energy
chain (cf. (2)), and from the point of development of the action itself, involving the absolute
construal. In the latter case one starts with the theme itself, the minimal action structure,
and by functionally elaborating the successive layers of such a relationship (by making
reference to causation and energy flow), one can arrive at the full thematic relationship.
Specifically, starting with the autonomous core of action and adding a conceptually dependent
layer of causation involving a participant which supplies energy, we can expand in this way
the description of the event, from the absolute autonomous construal to a fully fledged force-
dynamics dependent structure. The A/D asymmetry based on the functional layering so
defined can be represented by the following formula (Langacker 1991b: 292):
(15) (T) > (E1(T)) > (E2(E1(T))) > (E3(E2(E1(T))))) > (E4(E3(E2(E1(T)))))))

T represents a thematic relationship, i.e. a role such as theme (which can be represented by the
thematic role of zero, mover, patient, experiencer, and E which stands for the energy
responsible for the process. The A/D asymmetry involved in this layering can best be captured
by quoting from Langacker (1991b: 292):

Thus, if T represents a thematic relationship, and E the input of energy responsible for
the occurrence of a process, the progressive assembly of the complex event conception
in (a) The ice cracked; (b) A rock cracked the ice; (c) A waiter cracked the ice with a
rock; (d) The manager made a waiter crack the ice with a rock; (e) The owner had the
manager make a waiter crack the ice with a rock can be represented as follows: (T) >
(E1(T)) > (E2(E1(T))) > E3(E2(E1(T)))) > (E4(E3(E2(E1(T))))). The parentheses indicate
A/D organization, but if one ignores them and reads each formula linearly, it is
equivalent to an action chain (the initial formula, T, represents the degenerate case of an
action chain with a tail but no head.

Consider now the following sentences from English:

(16) a. The dry moss caught fire.

b. A piece of broken glass set the dry moss on fire.
c. Mark used a piece of broken glass to set fire to the dry moss.
d. Peter persuaded Mark to set fire to the dry moss with a piece of broken glass.

An event whose functional structure is elaborated to the greatest degree (i.e. E4(E3(E2
(E1(T))))) is exemplified by (16d) and (16c), while an event described in absolute (or zero)
terms (which involves only (T)), is given by (16a). By adding successive layers of causation
to the event portrayed as the absolute construal (cf. (16a)), the structures of the events
involved become more elaborate which, as already remarked, enables one to describe the
events in terms of force dynamics. One might say that structures such as (16c) and (16d)
which are arrived at in this way (i.e. by ascending via causation the energetic chain)
meet those descending the detransitivization scale (but see below).
Consider now the following dynamicity hierarchy for Polish middles (cf. Kardela
2000: 195)

(17) The dynamicity hierarchy for Polish middles

(i) no energy input (or: zero) > (ii) no energy control > (iii) energy control > (iv)
counter-to expectation energy impact
(18) (i) Jan drzema John was taking a nap > (ii) Jan zdrzemn si John dozed off; Jan
polizgn si na mokrej pododze i upad John slipped on the wet floor and fell down >
(iii) Jan przeszed si po Starym Miecie wstpujc do wspaniaych kociow John took
a walk round the Old Town visiting the superb churches (Lit. going into the superb
churches); Jan wlizgn si do pokoju John sneaked into the room > (iv) Wczoraj Jan
przeszed samego siebie Yesterday John surpassed himself [lit. John went beyond
himself].

(18i) represents cases of an absolute construal, with no energy input. The activity of Johns
taking a nap is described without taking into account all the circumstances accompanying this
activity. In (18ii) John does not have control over his activities: he is dozing off or slipping on
the wet floor regardless of his will. (18iii) involves construals which suggest some energy
control: John wilfully takes a walk round the city, visiting the churches. Finally, in (18iv),
which involves the sequence of the emphatic sam and the heavy reflexive siebie, the activity
depicted leads to a counter-to-expectation change: (18v) could, for instance, be said of a
person (here John) who, even to his own surprise, performed very well at the concert, playing
the violin.
We can present diagrammatically now the absolute construal relationship involving
Polish si-constructions as follows (cf. Kardela 2000: 198):
(19)

canonical event structure

Jan umy samochd

John washed the car.
.
S DO .

Reciprocals

Chopcy pobili si/jeden drugiego

Specified energy The boys beat each other
impact
S DO

reflexives

Jan umy si.

John washed himself.

S/DO

middle
Jan przeszed samego siebie
counter-to-expectation John surpassed himself.
energy impact +> Wanna si przelewa
The bathtub is overflowing
S/TH

energy control Jan wlizgn si do pokoju.

<+ John sneaked into the room
> Jan przeszed si po miecie
S /TH John walked around the city

no energy control Jan zdrzemn si

S /TH John dozed off.

absolute (zero)

Jan drzema.
no energy input John was taking a nap
S/TH
The diagram represents the energetic path of the absolute construal relationship. The path
starts with absolute constructions such as drzema taking a nap, i go, etc, incorporates
the middle portion of the absolute construal (inner box), representing the energy input
supplied by reflexives (including heavy reflexives which mark changes in the patient), and
proceeds, via reciprocals, to prototypical transitive constructions which profile the whole
energetic chain.
By combining now the insights given in (15) through (19), we obtain the following
results as defined by the functional A/D layering. The participant of drzema take a nap in
(18i) represents a thematic relationship T (i.e. the participant of drzema is a theme). The
addition of (non-controlled) energy in (18ii) makes the semantic (causative) structure of
zdrzemn si doze off dependent on drzema. This is so because causation is conceptually
dependent, since, as Langacker puts it, it makes inherent (albeit schematic) reference to the
change induced (Langacker 1991b: 287)). The structure of przej si (18iii), in turn, is
autonomous vis--vis the element of causation appearing in (18iv). As already mentioned, the
process of energy increase in (18iv) is so high that it produces a counter-to-expectation result.
Notice also that while the verbs enclosed in the box labelled as middle and a
reflexive si in umy si involve a single participant undergoing some change in the process
of dynamicisation, the events coded by the verb pobi si/pobi jeden drugiego beat each
other (reciprocal) and umy (co/kogo) wash sth/sb) involve two participants, one of
themin the case of the reciprocalbeing expressed by the reflexive si. In the last two
cases the external causers of the events of seeing and washing respectively are specified: it is
one of the participants who causes the other participant to see him and to wash it. The
most prototypical events are thus represented by expressions such as umy si and pobi
si/pobi jeden drugiego, which require two participants, one affecting the other by
transmitting the energy onto him. The upper figure represents then the flow of energy from
the participant coded by the subject of the sentence to the participant coded as an object. In
these two last cases the degree of dynamicity is the greatest of all the examples discussed so
far. These two last cases then, and especially the sentence Jan umy samochd prototypically
represent the canonical event model rebours, in which the energy flows from the participant
heading the chain to the participant which ends the chain.
The obvious question to ask now is why we need to reconstruct the energy chain
from an up-stream flow of energy perspective? In other words, why do need to look at an
event from both the energy flow perspective and the absolute construal point of view? To
answer this question we have to look first at middle voice constructions.
 According to Renata Grzegorczykowa, the use of the middle voice si is one of the most
important strategies of de-angentivization in Polish besides the passive voice and impersonal
constructions. Thus si appears in middles which are derived both from structures containing
imperfective (viz. (20)-(21)) and perfective (cf. (23)-(25)) verbs (Grzegorczykowa 1996: 59-
60):
(20) Te pyszne ciasteczka zjadaj si bardzo atwo.
These delicious cakes can be eaten very easily.
Lit. These delicious cakes eat very easily.
(21) Te piguki poykaj si bardzo atwo.
These pills can be swallowed very easily.
(22) Szkoa wyremontowaa si.
The school was renovated.
(23) Wszystkie ciasteczka si zjady.
All the cakes were eaten.
(24) Cay sok si wypi.
All the juice got/was drunk up.
(25) Koszule upray si.
The shirts were washed clean.

There are, however, certain restrictions on the formation of middles. Consider:

(26)? Ksiki wypoyczaj si (atwo, bez trudu, wasnie).

The books can be borrowed (easily, without difficulty, right now).
(27) ?Te towary kupuj si dobrze, atwo.
These commodities buy well, easily.
(28) *Wszystkie ksiki kupiy si.
All the books were bought.
(29) *Caa ksika si przeczytaa .
The whole book was read.

The very low acceptability of sentences such as (26) and (27) should, according to
Grzegorczykowa, be attributed to the importance of the agent who is normally engaged in the
activities designated by the respective verbs appearing in these sentences. These sentences are
not well-formed because they require the stronger role of an agent. Particularly interesting
are sentences such as (28) and (29), in which the absence of the stronger role of an agent is
even more glaring. Notice, however, that, when slightly modified to include the emphatic sam
and uttered with the intention of producing ironical statements, these sentences appear to be
well-formed:

(30) Te wszystkie ksiki same si kupiy, co?

Well, all these books got bought all by themselves, did they?
 [i.e. Somebody had to pay for all these books, dont you think? And it is me who paid for
them]
(31) Ta ksika sama si nie przeczyta.
This book will not get read all by itself .
[Somebody has to read it. And it will probably be you]

(30) and (31) are very important for two reasons: firstly, because they code events that happen
contrary to our expectations (hence the ironic tone of the utterance), and secondly, the
presence of a reflexive si + the marker of perfectivity + the emphatic sam mark the presence
of some (covert) causer of the events, which gain in dynamicity.
Yet, as observed by Grzegorczykowa, not all structures involving si can be treated as
de-agentivized constructions (pp. 62-63) (adapted):

(32) Drzwi otworzyy si/zamkny si. (cf. Jan otworzy/zamkn drzwi.)

The door opened (closed). (cf. John opened/closed the door.)
(33) Telewizor zepsu si. (cf. Jan zepsu telewizor.)
The TV set broke down. (cf. John damaged the TV set).
(34) Mysz zapaa si. (cf. Kot zapa mysz.)
The mouse got trapped. (cf. The cat caught the mouse.)
(35) Piotr uderzy si. (cf. Jan uderzy Piotra.)
Peter hit himself. (cf. John hurt Peter.)
(36) Sukienka zabrudzia si (cf. Marysia zabrudzia sukienk.)
The skirt got dirty. (cf. Mary made the skirt dirty.)

(32) - (33) can only be treated as non-prototypical de-agentivized constructions. This is so

because, as noted by Grzegorczykowa, de-agentivised structures are inherently ambiguous:
they may be the result of sbs activity or they can be the result of a self-incurred process, as
in Drzwi otworzyy si na skutek przecigu The door opened as a result of the draught, or
Telewizor zepsu si samoistnie The TV set broke down by itself. Nor can (34) be treated as
a de-agentivized construction derived from the sentence Kot zapa mysz The cat caught the
mouse because the mouse may have fallen prey all by itself to a mouse trap. Although (32)-
(33) (and perhaps (34)) can be treated as some sort of non-prototypical de-agentivized
constructions, Grzegorczykowa denies the status of de-agentivized constructions to sentences
such as (35) and (36), because in (35) the putative object is animate (it is a sentient person,
viz. Piotr), and (36) does not mean that somebody has made the skirt dirty on purpose. Let
us keep this assertion in mind, as this is one of the reasons why we need to look at events
from two perspectives: from the energetic chain perspective and the absolute construal
viewpoint.
 Another reason for this dual perspective on event structure follows from the behaviour
of si in the context of causality. Thus, when discussing verbs denoting psychological states
such as cieszy si cieszy enjoy oneself make sb. glad, dziwi si dziwi be surprised
surprise sb, gniewa si gniewa get angry (make) sb. angry, Grzegorczykowa offers
the following comment (translation is mine; p. 63):

In this case we have no doubt that in the pair of verbs, the causative construction is
secondary and is derived when the postfix si is dropped. We are dealing here not with de-
agentivization, but rather with the causativization of the processual verb. This is so
because state-processual verbs of the type martwi si worry, cieszy si rejoice, etc.
denote human emotional states caused by some event or situation. The use of a causative
verb indicates that some state will be induced, whereby a particular situation or an event,
but not a person, will be the causer of this state (as is the case with classic verbs of action),
e.g. The mother is worried about her son (about her sons behaviour)
Her son (her
sons behaviour) worries the mother, that is, the sons behaviour makes the mother
worried.
.
Grzegorczykowas insight is crucial here: on the one hand, the presence of si marks
the process of de-agentivization (or detransitivization, in general), on the other hand, the
presence of si with verbs such as cieszy si, martwi si, etc, signals (latent) causativity.
When si is dropped, we get the causative reading. It is clear then that, ifas it appears to
be the casesi plays an important role not only in detransitivisation but also in causation,
then our analysis of the absolute construal involving A/D layering (cf. (15)) must include si
as well.
Consider the following examples involving the heavy reflexive sobie and the light
reflexive si:

(37) a. Marysia ufarbowaa sobie wosy na rowo.

Mary has dyed her hair pink.
b. Marysia sama sobie ufarbowaa wosy na rowo.
Mary has dyed her hair pink all by herself.
c. Marysia kazaa sobie ufarbowa wosy na rowo.
Mary had her hair dyed pink.
(38) a. Zakr kurek woda si przelewa.
Turn off the tap the water is overflowing.
b. Zakr kurek wanna si przelewa!
Turn off the tap the bathtub is overflowing!
c. Zakr kurek nie pozwl wodzie przelewa si przez wierzch wanny!
Turn off the tap Do not let the water flow over the rim of the tub!

(37a) is ambiguous as it can either mean (37b), i.e. that Mary did the dying herself without
going to a hairdressers, or (37c), in which case the hairdresser did the actual dying for Mary.
Observe now that in the first case we have an energetic chain construal, where the reflexive
sobie marks the departure of the transitive construction from its prototype. The reflexive here
functions as a de-agentivizer (detransitivizer). On the second reading, (37a) must be treated as
a causative construction, with the hairdresser being the causer of the dying. Here the reflexive
functions as a causitivizer. It is obvious then that these two ways of looking at the event of
dying are not equivalent: in (37b) Marysia is an agent and still performs the activity, while in
(37b) the causer of this activity is the hairdresser, not Marysia (see Babby (1993) for a
discussion of similar examples in Russian).
Turning to (38), observe that (38a) has an energetic construal: woda the water is a
patient/theme. Because only this role is profiled here, woda is coded as the subject of the
sentence. The addition of si in (38b) changes the whole configuration: the verb becomes
transitivized and wanna the tub, not woda, becomes its subject. More importantly, this
means that (38b) could never be derived via the energetic construal because on this construal,
only an agent or instrument can belong to the source domain and thus induce changes in the
patient. Finally, (38c) is a case of a typical causative construction with a typical causative
verb pozwoli (przelewa si) let (it overflow).
Following Maldonados (1993) analysis of dynamic construals, we would like to claim
then that the function of si in constructions such as wanna si przelewa the bathtub
overflows is to impart dynamicity and unexpectedness to the absolute construal of a scene.
The event thus described acquires a quality of accidentality and, as Maldonado puts, it goes
counter to the expected conceptualization. (See the discussion in Maldonado (p. 551) of
examples such as El tejado se llovio The ceiling got rained through [lit. The ceiling rained],
La banera se salio Watter dripped out of the bathtub [Lit. The bathtub went out]; on the
dynamicising property of se in Spanish, see also Nishida 1994).
In conclusion, in this paper we have argued for the importance of looking at an event
from two perspectives: from the point of view of force dynamics and from the standpoint of
the absolute construal. We have adduced evidence for these two perspectives, claiming they
are not identical. We have also pointed to the Janus-like property of the Polish reflexive si.
On the one handseen from the force dynamics perspectivesi signals the de-
agentivisation process, while on the other hand, the same particleviewed from the absolute
construal perspectivebecomes a sign of the externally-induced causativisation process. The
ensuing event dynamicisation process is based on the A(utonomy)/ D(ependence)
functional layering as proposed by Ronald Langacker:
References
Grzegorczykowa, R. 1996. Reguy deantywizacji w polszczynie, Jazyk a jego uzivani.
Sbornik k jubileu O. Ulicnego. Praha.
Kardela, H. (2000) Dimensions and Parameters in Grammar. Studies on A/D Asymmetries
and Subjectivity Relations in Polish, Maria Curie Skodowska University, Lublin.
Langacker, R. (1991a) Concept, Image and Symbol. The Cognitive Basis of Grammar.
Mouton de Gruyter, Berlin.
Langacker, R. (1991b) Foundations of Cognitive Grammar. Vol 2. Descriptive Application.
Stanford University Press, Stanford, California.
Maldonado, R. (1992) Middle Voice. The Case of Spanish SE. Ph.D. Dissertation,
University of California, San Diego.
Maldonado, R. (1993) Dynamic Construals in Spanish, Studi Italiani di Linguistica Teorica
e Applicata 22, 531-566.
Nishida, C. (1994) The Spanish Reflexive Clitic se as an Aspectual Class Marker,
Linguistics 32, 425-458.