The relation between iconicity and
subjectification in Portuguese
complementation: Complements of
perception and causation verbs
RAINER VESTERINEN*
Abstract
The present paper examines the variation between finite and infinitive comple-ments of the Portuguese perception/causation verbs ver (‘see’), ouvir (‘hear’), sentir (‘feel’), deixar (‘let’) and fazer (‘make’) from a cognitive grammar per-spective. It is argued that the distribution of the structures main verb?+ finite /infinitive complement can be explained by iconicity and subjectifica-tion. The hypothesis is put forward that the structure perception verb?+ infinitive complement designates direct physical perception, while the structure perception verb?+?finite complement designates an inferential relation between the main verb and the complement event. In addition, the structure causation verb?+?infinitive complement designates direct cau-sation, whereas causation verb?+?finite complement designates an indi-rect causation with inferential features. Further, it is claimed that the inferen-tial and conceptually more complex character found in the structure main verb?+?finite complement represents a prime example of subjectification. Being so, it is argued that Portuguese complementation provides a remarkable connection between iconicity and s u bjectification.
Keywords: causation; cognitive grammar; complementation; iconicity; in-ference; perception; Portuguese; subjectification.
1. Introduction
Since Haiman’s (1980, 1985) seminal work on iconicity in language, func-tional and cognitive linguistics has shown a major concern in establishing a relation between iconic principles, on the one hand, and different complement
* A ddress for correspondence: Department of Spanish, Portuguese and Latin American studies, Stockholm University, Universitetsv?gen 10b, 106 91, Stockholm, Sweden. E-mail: rainer@isp.
su.se
Cognitive Linguistics 21–3 (2010), 573–600 0936–5907/10/0021–0573 DOI 10.1515/COGL.2010.019 ? Walter de Gruyter
574 R. Vesterinen
structures, on the other. A central claim holds that there is a correspondence between formal and conceptual complexity in the complement structure. An-other claim is that a greater formal distance between the main verb and the complement verb matches a greater conceptual distance between the events described by these verbs (cf. Achard 2000, 2002; Givón 1993, 2001; Maldo-nado and Nava 2002; Verspoor 2000).1
The aim of the current paper is to corroborate these claims, supplying evi-dence from the Portuguese language. In view of natural data, I will argue that the distribution of finite and infinitive complement structures of perception and causation verbs in Portuguese is highly motivated by conceptual differences, i.e., the conceptual distinction made between direct and indirect perception/ causation. Furthermore, I will suggest that these conceptual differences entail a higher degree of subjectification in the structure main verb?+?finite com-plement than in the structure main verb?+?infinitive complement. In this sense, the present paper aims at drawing attention to a plausible connection between iconicity and subjectification in finite and infinitive complement struc-tures. The notion of subjectification will be considered both from Traugott’s (1989, 1995, 1996) and Langacker’s (1990, 1999, 2003, 2006) frameworks. The Portuguese perception verbs ver(‘see’), ouvir(‘hear’) and sentir (‘feel’), as well as the causation verbs deixar (‘let’) and fazer (‘make’), share a remarkable feature, namely that of allowing both infinitive and finite comple-ment structures. Additionally, the infinitive complement structures of these verbs are divided into three different structures, depending on the position of the “logical subject” of the infinitive and on the use of the plain or the inflected infinitive.2 This produces four different complement structures, illustrated in (1– 4) with the causation verbs deixar and fazer:
(1) Os pais deixam/?f azem brincar os meninos (?V V?)
The parents let/make-pres: play-inf. the children
3p.p
‘The parents let/make the children play’
(2) Os pais deixam/?f azem os meninos brincar (?V OV?)
The parents let/make-pres: the children play-inf.
3p.p
‘The parents let/make the children play’
1. The term event and its different uses in the current paper should be understood in a broad sense
of the word. The term will be used to designate an action, a process or a state (cf. Silva 2004).
2. The Portuguese inflected infinitive is a typical feature of the Portuguese language. The in-
flected infinitive agrees with the subject in person and number and is formed by adding a suffixal subject morpheme to the plain infinitive in the following way: -? (1p.s.), -es (2p.s), -?(3p.s), -mos (1p.p), -des (2p.p), -em (3 p.p).
Iconicity and subjectification in Portuguese complementation575 (3) Os pais deixam/?f azem os meninos brincarem (?V SV?)
The parents let/make-pres: the children play-inf: 3p.p
3p.p
‘The parents let/make the children play’
(4) Os pais deixam que / fazem com que os
The parents let-pres: that / make-pres: with that the
3p.p3p.p
meninos brinquem (finite complement)
children play-pres.subj: 3p.p
‘The parents let/make the children play’
The infinitive structures in (1–3) are frequently referred to as VV, VOV and VSV (cf. Silva 2004, 2005). In the first example (1), the main verb (deixar/?f azer) is followed immediately by the plain infinitive, forming the VV structure in which the logical subject of the infinitive has a final position. In (2), the logical subject is inserted between the main verb and the plain infini-tive. Nonetheless, the use of a plain infinitive, without any subject agreement, produces the VOV structure. In other words, the logical subject of the infinitive is considered as the grammatical object. Finally, in (3), the word order is the same but there is subject agreement on the infinitive, which results in the VSV structure.
Turning to the structure main verb?+?finite complement, illustrated in ex-ample (4), it is easy to verify some formal differences. Whereas the infinitive complements follow immediately after the main verb, the finite complements are normally introduced by the complementizer que (‘that’). This is the case with complements of the perception verbs ver, ouvir and sentir, and with finite complements of the causation verb deixar. The finite complements of fazer diverge from this pattern, being introduced by the preposition com (‘with’) im-mediately followed by the complementizer. Another formal difference between the finite complements of perception and causation verbs is that the perception verbs take an indicative verb complement, while the causation verbs take a verbal complement in the subjunctive mood.3
Studies on complementation in Portuguese have traditionally been carried out from a generative approach to language. The main purpose of these studies has been to describe the grammatical contexts that allow infinitive complement structures; and the structure with the inflected infinitive in particular. In doing
3. Although the contrast between indicative and subjunctive complements is highly relevant in
relation to the perception/causation verbs, it goes beyond the scope of the present paper. See Maldonado (1995) and Vesterinen (2006, 2007a) for detailed analyses on the semantics of the subjunctive from a cognitive grammar perspective of language.
576 R. Vesterinen
so, the distribution of infinitive complements has frequently been explained by the Chomskyan theory of Government-Binding (cf. Brito 1995; Caetano et al. 1994; Raposo 1987). Needless to say, the semantic concern has been highly ignored within these frameworks. However, one exception to this tendency is found in Silva (2004, 2005). Departing from a cognitive grammar perspective (Langacker 1987, 1991), Silva distinguishes a continuum in which the initial position of the logical infinitival subject and the inflection of the infinitive render the event described in the complement more independent from the main clause event (cf. Silva 2004, 2005).
The aim of the present study is to go one step further. I will suggest that the structures main verb?+?finite/infinitive complement represent a prime example of iconicity. Thus, it will be shown that the structure perception verb?+?infinitive complement denotes a prototypical direct physical percep-tion, while the structure with a finite complement implies there to be an infer-ential relation between the perception verb and the complement. In the same manner, the infinitive complements of causation verbs tend to designate a pro-totypical direct causation, while the finite complements designate indirect cau-sation. It will also be argued that the indirect causation attested in the finite complements often is of an inferential type. In sum, the structure perception/ causation verb?+?finite complement provides evidence of the speaker’s (the conceptualizer’s) inferences about the events described by the main and the complement clause.
As indicated above, the inferential character found in perception/ causation verb?+?finite complement will be explained by the notion of sub-jectification. From Traugott’s perspective (1989, 1995, 1996), I will show that the meaning of the finite structures is utterly based on the internal (evaluative, perceptual, cognitive) described situation; and further, I will propose that the finite structure designates the speaker’s subjective belief state toward the prop-osition. From Langacker’s perspective (1990, 1999, 2003, 2006), I will argue that the inferential character found in perception/causation verb?+?finite complement subsumes semantic bleaching and a change in perspective and locus of activity, i.e., from an active object of conception to the mental scan-ning of a subjectively construed conceptualizer. Thus,the analysis in the pres-ent paper provides an alternative interpretation to the claim that finite comple-ments of causation verbs constitute a peripheral subcategory mainly expressing “purely” causal relations (cf. Verhagen 2005). In fact, the finite complements of perception and causation verbs seem to exhibit a striking parallelism regard-ing the notion of subjectification.
In the light of this analysis, I will ultimately argue that the finite complement structures provide evidence of a connection between iconicity and subjectifica-tion. The first step in this process is that a higher degree of indirectness be-tween the main event and the complement event in the structure main verb?+
Iconicity and subjectification in Portuguese complementation577 finite complement designates a more inferential relationship between these two events than does the structure main verb?+?infinitive complement. The second step in this process is that the inferential relation in the structure main verb?+?finite complement evokes the conceptualizer to a higher degree than is the case in the structure main verb?+?infinitive complement.
The outline of the paper is as follows: The results of a quantitative analysis of complement patterns regarding direct and indirect perception/causation are presented in section 2. Sections 3 and 4 are dedicated to a qualitative anal-ysis of the structures perception verb?+?complement and causation verb?+ complement respectively. Section 5 describes the relationship between ico-nicity and subjectification. The conclusions are presented in section 6.
2. Complement patterns: direct and indirect perception/causation
The linguistic material analysed consists of 600 natural examples extracted randomly from the Portuguese corpus Linguateca and from Portuguese I n ternet sites. The structure main verb?+?complement consists of 300 occurrences of the structure perception?verb?+?complement and 300 occurrences of the structure causation?verb?+?complement. Each verb type is further repre-sented with 150 finite and 150 infinitive cases. That is, the category p e rception?verbs consists of 150 finite complements and 150 infinitive complements, and the same holds for the category causation?verbs.
In order to make a distinction between direct and indirect causation/ perception, the following parameters were primarily considered: (1) direct per-ceptual experience of an event vs. inferential processes based on perceptual evidences, and (2) direct physical causation detectable in the outside world vs. causation as a mental experience. Regarding the perception verbs, the distinc-tion is noteworthy in cases like: I saw the children playing football vs. I see that you are tired. In the latter case, the conceptualizer does not actually see the event of someone being tired, but infers it on the basis of perceptual evidences. Hence, this case designates indirect perception. Likewise, mental causation in which the causer induces the causee to perform a certain action is considered to be a case of indirect causation. On the other hand, the direct manipulation of an object often refers to direct causation. Obviously, the cases discussed above are prototypical instances of direct and indirect perception/causation. The dis-tinction drawn between direct and indirect perception/causation will be dis-cussed in more detail in the qualitative analysis.
The quantitative analysis confirms a strong pattern in which the structure main?verb?+?infinitive?complement designates a direct perception/c a usation, while the structure main?verb?+?finite?complement designates a more indi-rect and inferential relation between the main verb event and the complement event. Accordingly, the structure perception?verb?+?infinitive?complement
578 R. Vesterinen
presents 130 cases of direct perception and only 20 cases of indirect (inferen-tial) perception. In contrast, the structure perception verb +?finite comple-ment displays a completely different pattern: 123 occurrences designate indi-rect (inferential) perception and 27 cases are of a more direct character. The association between complement type and direct /indirect perception is highly significant, χ2 = 139.02, df = 1, p < .0001 (two-tailed). The results of the struc-ture perception verb +?complement are shown in Table 1.
Turning to the structure causation verb +?complement , it is noteworthy that the same pattern prevails. The structure causation verb +?infinitive complement designates a more direct kind of causation (116 cases), while it only exhibits 34 cases of indirect causation. In contrast, the structure causa-tion verb +?finite complement provides a totally different tendency. In the latter, 122 cases designate indirect causation and no more than 28 cases desig-nate direct causation. The association between complement type and direct /indirect causation is statistically significant, χ2 = 101.08, df = 1, p < .0001 (two-tailed). The results of the structure causation verb +?complement are shown in Table 2.
It is noteworthy that the finite cases diverging from the general pattern often designate a kind of categorisation, e.g., I saw that the car was white . Another type that diverges from the general pattern corresponds to evidentials with the verb ouvir (‘to hear’) of the type: I heard that Peter is in town . In both cases, however, one could easily argue that the finite structures display indirect fea-tures. In the first case, the categorization of an entity entails a more elaborated mental process than the mere registration of an event; and the second case cor-responds to an information chain that certainly is indirect in nature. However,
Table 2. C ausation verbs + finite and infinitive complements.
Main Vcausation +
Finite compl.
Main Vcausation + Infinitive compl.Totals Direct causation Indirect causation Totals
28122150
116 34150
144156300
Table 1. P erception verbs + finite and infinitive complements.
Main Vperception +
Finite compl.
Main Vperception + Infinitive compl.Totals Direct perception Indirect perception Totals
27123150
130 20150
157143300
Iconicity and subjectification in Portuguese complementation579 they do not designate strictly inferential structures and, therefore, they are con-sidered direct cases in the present paper.
The infinitive cases with an indirect meaning, on the other hand, often correspond to the VSV structure with a human subject in both main and com-plement clause, e.g., They heard the opponent provoke the Prime Minister. That is, an occurrence in which the subject of the perception verbs not only hears something, but also draws a conclusion on the basis of the perceptual experience. This latter tendency regarding the VSV structure is also visible in a small number of cases of inducive causation designated by the verb fazer (‘to make’).
Tables 1 and 2 statistically corroborate the hypothesis that the two structures differ substantially regarding their conceptual content. The structure main?verb?+?infinitive?complement tends strongly to designate a direct p e rception/ causation, while the structure main?verb?+?finite?complement designates a more indirect kind of perception/causation. In subsequent sections, I will anal-yse these conceptual differences more thoroughly.
3. Complements of the perception verbs ver, ouvir and sentir
The distinction made in the current paper between direct perception, on the one hand, and inferential structures, on the other, is based on the folk model of the mind and the notion that physical perception is an immediate and spontaneous phenomenon. That is, in spite of modern theories of perception, there is a wide-spread idea that what we see, hear or feel is caused directly by an external stimuli in the outside world. We hear or see something because it occurs in our environment; and if it occurs within our auditory or visual field we have only a limited possibility to avoid it. There is, so to speak, a general idea that per-ception is direct and uncontrollable. Inference (or inferential structures), on the contrary, is related to our capacity for reasoning and to mental processes that lead us to certain belief states about how the outside world is shaped. As a re-sult, inference is conceived of as more indirect and more controllable than perception (cf. D’Andrade 1987; Verhagen and Kemmer 1997).
Relating the distinction made above to infinitive and finite Portuguese com-plements of perception verbs, it is interesting to verify a rather clear-cut cor-relation. The structure main?verb?+?infinitive?complement designates a d irect perceptual relation between the main verb perceiver and the process described in the complement. That is to say, an external stimulus in the outside world causes the perceiver to see, hear or feel something directly. In contrast, the structure main?verb?+?finite?complement does not merely describe a physical perceptual relation between the main verb subject and the comple-ment event, but an inferential one. Thus, this structure describes inference,
580 R. Vesterinen
reasoning and beliefs about the outside world.4 The infinitive cases in (5–7) illustrate the direct perceptual relation between the main verb subject and the event described in the complement:
(5) Em apenas vinte minutos que estive no café
In only twenty minutes that be-past:?in the coffee bar
1p.s
vi duas senhoras serem atacadas.
see-past: two ladies be-inf: attacked.
1p.s 3p.p
‘I was only in the coffee bar for 20 minutes, but I saw two ladies being attacked.’ [Diário de Aveiro-N2240-1]
(6) O repórter do Diário de Coimbra que na altura
The reporter of Diário de Coimbra that in the moment
se deslocou ao local ouviu populares
refl. move-past: to the place hear-past: people
3p.s 3p.s
imputarem a prática do crime a alguém.
attribute-inf: the practice of the crime to someone.
3.p.p
‘The reporter of Diário de Coimbra, who at that moment was moving
towards the crime scene, heard people attribute the crime to someone.’ [Diário de Coimbra-N0859-1]
(7) “A meio da noite senti a minha cama
T o middle of the night feel-past: the my bed
3p.s
tremer, como um terramoto”, afirmou um dos
shake-inf: like an earthquake, affirm-past: one of the
3p.s, 3p.s
sobreviventes.
survivors.
‘In the middle of the night I felt my bed shaking like an earthquake, said one of the survivors.’ [Diário de Leiria-N0991-1]
In (5), the main verb subject, expressed by the first person singular preterit tense vi (‘I saw’), personally witnessed the act of two ladies being attacked. In the following example (6), the reporter in a direct manner heard (ouviu) some-one attribute a crime to someone. Finally, in (7), the main verb subject felt (senti) the bed shaking. Thus, the common denominator of these cases is a
4. See also Vesterinen (2007b) for a detailed examination on the matter.
Iconicity and subjectification in Portuguese complementation581
d irect perceptual relation between th
e main verb subject and the process d escribed in the infinitive complement.
Turning to the structure main?verb?+?finite?complement, it is difficult to find this kind of perceptual relation between the main verb subject and the complement event. Instead, this structure designates an inferential relation in which the main verb subject, basing him-/ h erself on some perceptual experi-ence, is drawn towards a conclusion about how the outside world is shaped. Examples (8–10) illustrate this phenomenon:
(8) Portanto, . . . fico sempre desolada quando vejo
Therefore . . . get-pres: always distressed when see-pres:
1p.s 1p.s que a comunica??o social dá a sensa??o de
that the mass media give-pres: the impression of
3p.s
estar controlada.
be-inf: controlled.
3p.s
‘Therefore . . . I always get distressed when I see that the mass media
give the impression of being controlled.’ [Diário de Aveiro-N1448-1]
(9) ouvem que a limpeza étnica e o apartheid
hear-pres: that the ethnic cleansing and the apartheid
3p.p
s?o a principal característica da política
be-pres: the typical trait of the politics
3p.p
de Israel.
of Israel.
‘They hear that ethnic cleansing and apartheid are typical traits of
Israelian politics.’ [?h ttp://https://www.doczj.com/doc/1c9572910.html,/luta.htm]
(10) Mas gostava de ver outras mulheres, bastante mais
But like-past: of see-inf: other women, much more 1p.s1p.s
novas na política, mas sinto que há um
younger in the politics, but feel-pres: that be-pres: a
1p.s 3p.s
grande desinteresse.
great disinterest.
‘But I would like to see other, much younger women, in politics, but I
feel that there is a great lack of interest.’ [Diário de Aveiro-N1448-1]
582 R. Vesterinen
The difference between these cases and the infinitive ones is that (8–10) de-scribe events that are not detected directly by the main subject’s perceptual apparatus. In other words, in (8) it is difficult to actually witness the event of mass media being controlled. Rather, the main verb subject draws a conclusion about the state of the mass media. In (9) the main verb subjects, expressed by the verb ouvem (‘they hear’), do not actually hear the event of ethnic cleansing and apartheid in Israelian politics, but something they heard made them realize that this is a typical trait of Israelian politics. And, in (10), the main verb sub-ject does not feel the event of great disinterest in the same direct manner as in the infinitive example (7). In sum, these cases do not express a direct physical perceptual relation between the main verb subject and the complement event. Rather, they designate an inferential process.
Being so, a question to be asked is whether it is possible to insert an infini-tive complement in the cases studied above and still have an inferential rela- tion between the main verb subject and the complement event. This problem may be addressed by means of a substitution test that could verify if the finite cases in (8–10) can be modified into infinitive ones without any semantic consequences:
(11) ?Vejo a comunica??o social dar a sensa??o de estar controlada
* I see the mass media give-inf: 3p.s the impression of be-inf: 3p.s controlled’5
(12) ?Ouvem a limpeza étnica e o apartheid serem a principal característica
da política de Israel
* They hear ethnic cleansing and apartheid be-inf: 3p.p typical traits of Israelian politics’
(13) ?Sinto haver um grande desinteresse
* I feel be-inf: 3p.s a great disinterest
In fact, these examples seem to highlight the semantic differences between the finite and the infinitive complements. Whereas finite complements may ex-press thoughts and propositions about the world, including cases like (8–10) that are of a more stative character, the infinitive complement seems more ap-propriate to describe processes. This produces a semantic conflict in (11–13) where the insertion of an infinitive complement creates expectations of a direct physical perceptual relation between a perceiver and a perceived process (cf.
5. The English translations of the modified and fabricated examples are made in a word-to-word
order fashion in order to capture the original Portuguese structures. While they are not always grammatical, they are intended to sustain the semantic considerations under discussion.
Iconicity and subjectification in Portuguese complementation583 Perini 1977: 48–51; Vesterinen 2007b: 271–273).6This conflict is perhaps even better illustrated in the following examples:
(14) O Jorge vê que a Maria tem muito trabalho
George sees that Mary have-pres: 3p.s much work to do
(15) ?O Jorge vê a Maria ter muito trabalho
* George sees Mary have-inf: 3p.s much work to do
(16) O Jorge vê que a Maria está cansada
Jorge sees that Mary be-pres: 3p.s tired
(17) ?O Jorge vê a Maria estar cansada
* George sees Mary be-inf: 3p.s tired
(18) O Jorge vê que a Maria é advogada
George sees that Mary be-pres: 3p.s a lawyer
(19) ?O Jorge vê a Maria ser advogada
* George sees Mary be-inf: 3p.s a lawyer
Once again, the peculiarity of the construction with infinitives (15, 17, 19) is related to the fact that propositions like she has much work to do, she is tired and she is a lawyer are not so easily detected in the outside world. On the con-trary, the main verb subject infers these states from some prior perceptual evi-dence. Therefore, the finite complements can easily be followed by a causal clause, introduced by porque (‘because’), that explains why the main verb sub-ject infers the complement event, e.g., vejo que estás cansado porque bocejas constantemente (‘I see that you are tired because you are yawning constantly’). This is not the case with infinitive complements:
(20) ?Vi duas senhoras serem atacadas, porque um homem estava zangado
com elas.
* I saw two women be-inf: 3p.p attacked, because a man was angry at them
(21) ?Ouviu populares imputarem a prática do crime a alguém, porque ele
era o culpável.
* He heard people attribute-inf: 3p.p the crime to someone, because he was guilty
(22) ?Senti a minha cama tremer, porque houve um terramoto
* I felt my bed shake-inf: 3p.s, because there was an earthquake)
6. The grammaticality/acceptability judgements in the present paper are based on my intuitions.
As such, they are consistent with the approach used in Perini (1977) and Vesterinen (2007b, 2008).
584 R. Vesterinen
The reason why examples (20–22) seem a bit odd is probably because they express an inferential relation that explains the cause of the complement event and not the reason for the main subject’s perceptual experience. Thus, it may be true that the man in (20) attacked the women because he was angry at them, but this does not explain why the main verb subject saw the event. A more likely explanation would be that he saw the event because it occurred within his visual field. Likewise, (21) would be more felicitous if it explained the reason why the main verb subject heard someone say something and not why someone said it. Finally, it may be true that the bed in (22) shook because of the earthquake. However, the reason why the main verb subject felt it was that he was in bed when it shook. In sum, infinitive cases like (20–22) designate perceptual relations. Therefore, a causal clause explaining the perceptual event is more felicitous with the structure main?verb?+?infinitive?complement. The cases studied above (5–22) seem to highlight the conceptual difference between finite and infinitive complement structures. The infinitive c o mplement structures designate a prototypically direct sensorial perception in its relation to the main verb, whereas the finite complement structures designate an infer-ential relation. Nevertheless, a remaining question is how to explain this differ-ence. In other words, why do the infinitive structures designate direct percep-tion and the finite ones inference?
A reasonable answer to this question can be found in the notions of event integration(Givón 1993, 2001) and grounding(Langacker 1990). A funda-mental idea underlying the notion of event integration is that a complement that is formally simple, and is lacking morphological information, tends to be integrated into the event described by the main verb. Likewise, a more elabo-rated complement gains a certain independence in its relation with the main verb and may be conceptualized as an event on its own. This difference is shown in the following examples:
(23) A s pessoas vêem/ouvem/sentem que estas decis?es trazem benefícios no
futuro
People see/ h ear/feel that these decisions bring-pres: 3p.p benefits in the future
(24) ?As pessoas vêem/ouvem/sentem estas decis?es trazerem benefícios no
futuro
* People see/ h ear/feel these decisions bring-inf: 3p.p benefits in the future
The difference between these examples is that the finite complement in (23) may designate a future event while the main verb designates an event in the present. That is to say, the main verb event and the complement event can be conceptualized as two different events. This is not the case with the infinitive
Iconicity and subjectification in Portuguese complementation585 complement in (24). The temporal expression no futuro (‘in the future’) has the effect of dislocating both the main and the complement event to the future. In other words, the absence of tense in the infinitive complements makes it more dependent on the temporal profile of the main verb.7
Further, the difference between a finite and infinitive complement pertains to the effect of grounding the complement event so that it bears some relationship to the speech event, i.e., the ground. Thus, the absence of tense in the infinitive complement implies there to be no such relation. Instead, the event described in the complement is “reached” through the main verb. This difference is i llustrated in Figure 1:
Figure 1. T he grounding of the complement event.
As indicated in Figure 1a, there is no direct relation between the ground and the infinitive complement event. This situation is illustrated with the absence of an arrow between the ground and the complement event. Even so, there is a relation in the sense that the ground has access to the event through the main verb. Simultaneously, a process that can be detected in the outside world does not need to be grounded in the same way as a proposition about the outside world. Therefore, it is not surprising that the infinitive complement designates a perceptual event, i.e., the physical perception of the main verb subject. On the other hand, the grounding of the complement event has the effect of locat-ing the event in relation to the speech act participants’ conception of reality (cf. Langacker 1991, 2004, 2008). Thus, the finite structure does not actually
7. The higher degree of dependence of the infinitive complement may be even more salient, tak-
ing the following case into consideration:
?Eles viram/ouviram/sentiram estas decis?es trazerem benefícios no futuro
*They saw/ h eard/felt?/these decisions bring-inf: 3p.p benefits in the future
It is actually difficult to separate the complement event temporally from the main verb event without creating a sentence that seems extremely odd.
586 R. Vesterinen
designate the perception of the main verb subject. Rather, it is the ground (the speaker/conceptualizer) who expresses his propositional attitude about how the world is shaped.
Although the ground is always the primary conceptualizer of a linguistic expression, the situation described above reflects a shift in perspective and vantage point. The infinitive complement implies that the complement event is seen, heard or felt directly by the main verb subject. A finite complement clause, on the other hand, changes this viewing arrangement. Rather than de-scribing the perception of the main verb subject, the finite complement desig-nates the mental process of the ground (the primary conceptualizer). In this paper, this situation is considered to be a case of subjectification and will be further studied in Section 5. The following section, however, will focus on complements of causation verbs.
4. Complements of the causation verbs deixar and fazer
The notion that causation can be of a more direct or indirect nature is com-monly based on a distinction between a physical and more direct kind of causa-tion, on the one hand, and a mental and more indirect causation, on the other. Likewise, direct physical causation is often claimed to correlate with a situa-tion in which the causer has a high degree of control over the causee, while the opposite is true for mental and indirect causation of a more inducing, permis-sive or enabling kind (cf. Shibatani 1976; Kemmer and Verhagen 1994; Shibatani 2002; Shibatani and Pardeshi 2002). Accordingly, Shibatanti and Pardeshi (2002: 89) point out that direct causation prototypically reflects a causal relation with a patientive causee, whereas the causee is more agentive in indirect causation.
This claim is also made by Verhagen and Kemmer (1997) and Song and Wolff (2004). The former define indirect causation as “a situation that is con-ceptualised in such a way that it is recognized that some other force besides the initiator is the most immediate source of energy in the effected event” (?V erha-gen and Kemmer, 1997: 67). Additionally, Song and Wolff (2004) introduce the no-intervening criteria in which the causation is claimed to be direct if there is no intervening causer between the initial causer and the final patient. Another trait, discussed in Song and Wolff (2004), is the degree of intention of the caused event. If the causer intentionally brings about an event, the causa-tion is said to be direct. If the resulting event is non-intentional, on the other hand, the causation is understood of as indirect. This distinction is exemplified with the difference between “the girl broke the vase” and “the girl caused the vase to break”. In the former case, the caused event is claimed to be intended; and therefore it designates direct causation. The latter case, in contrast, desig-nates a non-intended causation, and correlates with indirect causation (cf. Song
Iconicity and subjectification in Portuguese complementation587 and Wolff, 2004: 240?–241). Figure 2 summarizes the distinctions discussed above:
Figure 2. D irect and indirect causation.
The idea underlying the distinction made in Figure 2 is that indirect causation is of a more mediated kind. For example, direct physical manipulation of an object, i.e., Joe threw away the stone designates an event in which there is no intervening force between the causer and the caused event. On the other hand, mental manipulation typically involves a mind-to-mind relation in which the causer tries to achieve the caused event by changing the world-view of the causee. And as Verhagen and Kemmer (1997: 71) state: “one cannot reach into another person’s mind and directly cause him or her to do, feel, or think some-thing”. In this case, the causee is of a more agentive kind, and responsible for his or her own actions.
However, it may be difficult to sustain that intention and control are relevant features in the distinction between direct and indirect causation. It is, for ex-ample, easy to imagine a physical causation in which one person unintention-ally pushes another person, and thereby causes him to fall. This would cer-tainly qualify as direct causation without being intentional or displaying a high degree of control over the caused event. Another problem arises with the en-abling or permissive mental causation. Verhagen and Kemmer (1997) seem to classify this causation as indirect “since the initiator may be considered the responsible for the consequences” (?V erhagen and Kemmer 1997: 68). The view taken in the present paper is that the enabling or permissive causation in many cases reflects a situation where the causee is more likely to be self pro-pelled into an action. Thus, the removal of a hindrance (social or physical) d irectly allows him to accomplish the action. This is, for example, the case in a sentence like: O pai deixa o seu filho sair(‘the father lets his son go o ut-inf. 3p.s’).
Being so, the distinctions made in Figure 2 are not to be seen as absolute, but as possible manifestations of a more basic parameter, elaborated by Shibatani
588 R. Vesterinen
and Pardeshi (2002). This parameter considers the spatiotemporal profile of the causal event structure as a whole, and the conceptualization of the situation as one single event or two separate ones. If there is a spatiotemporal overlap be-tween the causing event and the caused event?—?t hat is, where the causer’s ac-tivity and the caused event are not so easily distinguished from one and other?—?t he causation is considered to be of a direct kind. However, if the caused event has some degree of autonomy, being conceptualized as an event on its own, the causation is of a more indirect kind. This difference is shown in Figure 3:
Figure 3. D irect and indirect causation (cf. Shibatani and Pardeshi, 2002: 90).
To begin with, Figure 3 has the advantage of explaining the prototypical differ-ences between direct and indirect causation. The direct physical causation typically displays a temporal overlap between the causing and the caused event so that the two events are conceptualized as one. In a situation where someone pushes, kicks or hits something, for example, the causing event has immediate consequences, which lead to the conceptualization of a single event. The caus-ing and the caused events are not clearly distinguishable. In contrast, the men-tal inducive causation suggests there to be two separate events: the inducive and the action event. In other words, an agentive causee may decide to act after considering the causation event.8
Further, Figure 3 may account for the permissive and enabling side of causa-tion. If the causal relation is conceptualized as a single event, i.e., if the re-moval of a hindrance has some immediate consequences, it is more likely to be categorized as direct causation. The causing and the caused event share the
8. Shibatani and Pardeshi’s original figure (2002: 90) does not include the terms causer and
causee, but the semantic roles agent and patient. The reason for not using the original terms is that causation, in the view of the present paper, does not need to imply an agentive causer. In
a case like a stone in the garden made him stumble and fall, for example, it is not really con-
vincing to say that the stone is agentive, neither that the causee is fully patientive (cf. Vester-inen 2008).
Iconicity and subjectification in Portuguese complementation589 same spatiotemporal profile. If the causing and caused events have different spatiotemporal profiles, on the other hand, this implies a more indirect kind of causation. This distinction seems to be rather straightforward with the enabling/permissive verb deixar:
(25) As janelas abertas deixam entrar o cheiro
The windows open let-pres: enter-inf: the scent
3p.p 3p.s
do campo.
of the countryside.
‘The open windows let the scent of the countryside enter.’ [?h ttp://www.
quinta-de-s-lourenco.pt/]
(26) Em plena rua Ivens em Lisboa, uma jovem
In middle street Ivens, in Lisbon, a young girl
deixa cair uma pasta cheia de papéis em frente
let-pres: fall-inf: a folder full of papers in front
3p.s 3p.s
de outra jovem.
of another young girl.
‘In the middle of Ivens street, a young girl drops a folder full of paper
in front of another young girl.’ [?h ttp://www.seleccoes.pt/?R evista/
detalhe.asp?tipo=detalhe&área=16&ID=5632&Grup o=77]
Both (25) and (26) designate physical causal relations in which the causee has some intrinsic tendency towards motion. In (25), the scent of the countryside has a propensity for extension and, therefore, the open windows directly allow it to enter the room. The meaning of deixar in this context would be more or less not impede.9 Likewise, the causer in (26) looses the grip of the causee (the folder) and?—?o beying the laws of gravity?—?t he folder immediately starts fall-ing. In other words, there is a spatiotemporal overlap between the causing and the caused events. This situation certainly reflects a direct causation, desig-nated by the structure main?verb?+?infinitive?complement.
In fact, a finite complement in these examples would produce some s e mantic concerns. In (25), for example, the semantic considerations are highly related to a conflict between a non human causer and a more elaborated causation in-cluding mental processes and reasoning. Being so, a finite complement in (26) seems to designate a mental process. The modification of (25–26) into struc-tures with finite complements illustrates this difference:
9. See Silva (1999) for a detailed examination on the semantics of deixar. Silva distinguishes
three basic meanings of deixar with verbal complements: the permissive act of allowing something, the ending of an impediment and to not impede an event.
590 R. Vesterinen
(27) ?As janelas abertas deixam que entre o cheiro do campo
*The open windows let that enter-pres.subj: 3p.s the scent of the countryside
(28) U ma jovem deixa que uma pasta cheia de papéis caia em frente de outra
jovem
*A young girl lets that a folder full of paper drop-pres.subj: 3p.s: in front of another young girl
The use of a finite complement in (27) creates a sentence that, in fact, seems semantically strange. This strangeness is probably due to the relation be- tween finite complements and mental processes (cf. Vesterinen 2008). If (27) would designate a process in which a human causer opens a window in order to get some fresh air, the use of a finite complement would be expected. In the present case, however, neither the causer (the open windows), nor the causee (the scent) fulfils the qualification of being an entity capable of mental processes and reasoning. Therefore, the human causer in (28) is un-problematic, but the finite complement seems to change the semantics of the sentence drastically. The most plausible interpretation is that the causer (uma jovem) has the possibility to impede the folder from falling but decides not to do so.
The examples studied above (25–28) illustrate that direct causation with the permissive or/and enabling verb deixar matches the structure main?verb?+ infinitive?complement. The following examples (29–30) designate a more mental and indirect causation. Thus, the use of finite a complement is not surprising:
(29) às vezes as pessoas deixam que os mesmos
Sometimes the persons let-pres: that the same
3p.p
problemas as tornem infelizes por anos.
problems them make-pres.subj: unhappy for years.
3p.p
‘Sometimes people let the same problems make them unhappy for
years.’
(30) Como era uma pessoa simples e prestativa
Because be-past: a person modest and helpful
3p.s
deixava que na altura as pessoas levassem
let-past: that at the time the persons take-past.subj:
3p.s 3p.p
Iconicity and subjectification in Portuguese complementation591 da sua mercearia produtos por “fiado”.
from the his/ h er grocer’s shop products for trust-part.
‘Being a modest and helpful person he/she let people buy on credit in
his/ h er grocer’s shop.’ [?h ttp://https://www.doczj.com/doc/1c9572910.html,/albertino/ historial.
html]
The first case (29) designates a causal relation in which problems successively put the causee in a state of misery. This mental process is indeed of a more dif-fuse and indirect character than the direct physical causation attested in the cases with infinitive complements. Further, it involves the conceptualization of two different events: the emergence of a problem and the state of unhappi-ness. The following example (30) shares the feature of a more complex causal event structure. Indeed, knowing that there is a possibility to buy something on credit leads to the reconsideration of the offer and, further on, to the event of accepting?—?o r not accepting?—?t he offer. To summarize, a more diffuse and indirect causation is coded by the structure main?verb?+?finite?complement. The same pattern prevails regarding the causation verb fazer. Direct causa-tion, often of a physical kind, is designated by infinitive complement struc-tures, whereas mental and indirect causation seems to require the structure fazer + finite?complement. The following examples will suffice to illustrate this difference:
(31) Aos 33 minutos, Bruno num pontapé de ressaca
To the 33 minutes, Bruno in a kick of hangover
faz a bola sair rente ao poste.
make-pres: the ball go out-inf: close to the goal.
3p.s 3p.s
‘In the 33rd minute, Bruno just misses the goal with a hangover-like
toe-poke’ [?V iana Diário-N0901-1]
(32) O aumento do contacto com cenas violentas
The rise of the contact with scenes violent
faz com que a crian?a reaja
make-pres: with that the child react-pres.subj:
?3p.s 3p.s
mais tardiamente a pedir ajuda ou a intervir para
more later to ask-inf: help or to intervene-inf: to
3p.s 3p.s
apaziguar uma luta entre outras crian?as.
calm-inf: a fight between other children.
3p.s
‘The increased contact with violent scenes makes the child react later
when asking for help or intervening in order to calm down a fight
592 R. Vesterinen
between other children.’ [?h ttp://www.medicosdeportugal.iol.pt/
action/2/cnt_id/72/]
Thus, the physical direct causation of hitting a ball and thereby making it move in a certain direction in example (31) is designated by the structure fazer + infinitive?complement. Conversely, in the following example (32), the struc-ture fazer + finite?complement designates a mental and more indirect type of causation. Obviously, the examples also differ in spatiotemporal profile. In (31), it is difficult to distinguish the causing and caused event from one another. There is, so to speak, a temporal overlap between them. In (32), on the other hand, it is obvious that a prior event (the contact with violent scenes) is said to result in a later event (a certain loss of empathy in witnessing authentic violence).10
It is also worthy of notice that the structures with an infinitive complement tend to designate emotional events that are understood as being more direct and uncontrollable. In these cases, the causing event creates a direct feeling, which seems to enter the causee before the acts of mental processes and rea-soning (cf. D’Andrade 1987). The finite complements, on the contrary, desig-nate actions that are consequences of a prior feeling:
(33) O primeiro dente, o primeiro cocó e xixi no
The first tooth, the first poo and pee in the
bacio, as primeiras palavras, os primeiros passos e
diaper the first words, the first steps and
tudo o que faz os pais orgulharem-se
everything the that make-pres: the parents proud-inf:
3p.s 3p.p refl.
dos filhos.
of the children.
‘The first tooth, the first poo and pee in the diaper, the first words, the
first steps and everything that make the parents proud of their
children.’ [?h ttp://www.jornaldeleiria.pt/index.php?article=7186
&visual=2]
10. Another piece of support for a correspondence between finite and infinitive complements, on
the one hand, and direct and indirect causation, on the other, is found in Silva (2005). As Silva comments, examples like:
?O Jorge deixa que/f az com que a Maria parta neste momento
*George lets that?/makes with that Maria leave-pres.subj: 3p.s in this moment’
have a minor acceptance in Portuguese. This minor acceptance could easily be explained by the notion that an expression that signals temporal contiguity (‘neste momento’) is not compatible with a structure that designates two different events, i.e., the structure deixar/f azer + finite complement.