Middle-Subjunctive Links

        This paper explores the relationship between the clitic se and 
indicative/subjunctive mood choice in Spanish. The Spanish clitic se has
generally been treated as a marker of detransitivization. As opposed to previousanalyses, this paper shows that se also imposes an increase in transitivity. 
The participant's level of involvement is increased so that s/he can exploit to the maximal extent the properties of the direct object. Transitivity increase
is shown to be determinant for mood choice in the complement clause. The notion of dominion, as depicted by the Reference Point Model in Cognitive Grammar 
(Langacker Volume II in press), is proposed as a principled way of accounting
for the development of full exploitation middle constructions which determine 
the choice of indicative mood in the complement clause.



0. The problem

It is a well-known phenomenon in Spanish grammar that there is some type of interaction between indicative/subjunctive mood choice and use of the so-called 
"reflexive" pronoun me/te/se. Since in all the cases that I will address, se 
does not have a reflexive function (footnote 2), I will use the term middle to
refer to the value of this clitic. The details of the middle/mood interaction 
have not yet been addressed in the current literature. The phenomenon to be
observed is whether the use of middle se in the main clause determines the 
choice of indicative or subjunctive mood in the complement clause. In the
examples, I will use the abbreviations (IND) and (SUBJ) to mark the mood of 
the complement clause. The problem to be analyzed can most obviously be shown
in the following contrastive examples:

1.Me temo que la policia va a reprimir a los estudiantes (IND)
`I am afraid that the police will repress the students'
2.Temo que la policia vaya a reprimir a los estudiantes (SUBJ)
`I fear that the police may repress the students'

While the use of the middle marker me requires an indicative complement clause, as in (1), its absence allows the occurrence of a subjunctive complement, as 
in (2). Although the middle/indicative interdependence is clear, the possibility
of having a middle/subjunctive combination with a meaning close to the one in 
(2), but rather different from (1), is also available under very specific 
circumstances:

3.Me temo que la policia vaya a reprimir a los estudiantes (SUBJ)
`I fear that the police may repress the students'

The analysis of the example in (3) depends crucially on the proper 
characterization of the first two contrasting examples.  It is more striking
that the most common pattern in Spanish does not follow the pattern exposed for
temer/temerse. With a wide variety of verb types, the middle marker se 
determines exactly the opposite results in the complement clause. Instead of 
taking an indicative complement, as in (1), it favors the use of subjunctive 
mood:

4.b.Me alegro de que vengas (SUBJ)
4.b.??*Me alegro de que vienes (IND)
   'I'm glad that you are coming'

Examples like (4.b) depend strongly on context and are rather marginal (footnote3).  The middle/subjunctive combination given in (4.a) constitutes the
standard use.  The problem to be explained can be summarized in the following
manner: while in alegrarse and other verb types the use of the middle marker
favors a subjunctive complement clause, there is a less common pattern for 
verbs of the temer type where se determines indicative mood choice in the
complement clause.  I propose that the middle se marker systematically 
increases the level of activity of the crucial participant-- generally speaking the subject--and that depending on the meaning of the verb there can be two
very different representations of the experiencer: an energized undergoer and a full exploitation active participant. Full exploitation se corresponds to cases in which the subject is maximally active, as s/he fully exploits an object
located in his/her dominion. The level of activity of an energized undergoer is much more limited. While active membersdetermine indicative mood choice, less 
active ones favor subjunctive.  This paper is organized in the following manner.In section I, I propose the notion of dominion as a crucial one for the proper 
characterization of active and passive experiencers. In section II, the notion
of dominion is proposed as the proper mechanism to differentiate indicative fromsubjunctive mood. In section III, I analyze some notions of activation imposed 
by middle se. In the last section, I link these patterns with the use of 
indicative/subjunctive mood in Spanish. 


I. Dominion
The notion of dominion, as proposed by Langacker (Volume II, in press), is a
component of the Reference Point Model (RPM). Among other linguistic phenomena, the RPM is intended to capture a wide range of possessive relationships. Since 
most of the constructions in this paper involve some abstract kind of
possession, I will make a few remarks about this issue. It is well known that 
the linguistic category of possession does not reduce to a single familiar 
value, such as ownership. There is a collection of relationships that the thing possessed may hold with respect to the possessor: a part (my hands), a relative (your mother), an unowned possession (the baby's crib), something hosted (the 
cat's fleas), a situation (his predicament), a related action (Maura's 
misunderstanding), something that fulfills a particular function (my bus,
her school teacher), and  many other possible relationships. Among the wide
variety of proposals to account for possession, Seiler (1983) proposes,
based on cross-linguistic considerations, that "linguistic possession consists
in the relationship between a substance and another substance" where one of themis animate, human and Ego or close to the speaker.  According to this definitionthe coexistence of two elements in the same domain is sufficient to establish
a possessive relationship. While this characterization is abstract and general enough to handle all the data, it may be too general since it doesn't include a
principled manner to account for the fact that some asymmetries are observable 
in possessive relationships: the fact that the whole is generally construed as
the possessor of a part, and the fundamental fact that in the association of twoobjects the owner, not the owned element, is generally identified as the 
possessor. Langacker (Volume II, in press) has proposed a similarly abstract 
characterization which accounts for such asymmetries. Instead of merely 
involving the coincidence of two substances in the same domain, it is proposed
that possession is a relationship in which one entity is located 
(conceptualized) in relation to another. According to the Reference Point Model
some entities are more easily located with respect to others that are in nature more salient and perceivable.  A salient object serves as reference point to locate a non-salient object that lies within its vicinity. Each reference point 
anchors a region, a dominion in which both the reference point and the target 
may be located. All the cases of abstract possession mentioned above are
accommodated by the reference point model. The reference point is the possessor and the target is the entity possessed. The asymmetries noted previously
are seen as the consequence of construing a salient entity as the reference
point for locating a less salient one. A whole is thus the possessor of its
(body) parts because the whole is more prominent (Brigitte's legs not *legs' 
Brigitte). Owners are possessors because people are more likely to be recognizedindividually than their inanimate possessions (the boy's knife not *the knife's boy). Finally, the cat is the possessor of its fleas (the cat's fleas) not only because the cat is more prominent, but also because there is a high level
of empathy established with respect to the cat's experience. The Reference PointModel constitutes an abstract schema that captures the conceptualizer's 
processing of a wide variety of possessive relationships. This abstract model,
however, is based on prototypical cases where possession is established in a 
concrete sense. The salience asymmetries suggested above generally correspond toactual differences between the possessor and the possessed element in the 
objective event.  It is actually the case that, in prototypical cases of 
possession, owning involves a participant establishing concrete control over a
specific object located in her/his dominion. Consequently the possessor is
conceptualized as actively interacting with a non-active element. By saying mi
casa `my house', mis ideas `my ideas', casa and ideas are objects at the 
possessor's disposal susceptible of being manipulated for different purposes. 
Since these objects are in my dominion I can choose to impose changes on 
them (trade them, destroy them, improve them and so forth). In a similar manner the possessor is active in the sense the s/he experiences mental or emotional 
sensations with respect to objects located in her/his dominion.  Body parts
(mis manos `my hands' mis ojos `my eyes') not only can be manipulated at the
possessor's will, but also the changes they may undergo imply an experiential 
activity by the possessor (Mis ojos estan irritados y no puedo ver `My eyes
are irritated and I can't see', Las pulgas del gato no lo dejan en paz `The 
cat's fleas don't leave it alone' [Lit:leave it in peace]). The last example is crucial for the correct understanding of these asymmetries. The cat does not 
possess the fleas crawling on it. It is only the case that the presence of fleasin the cats dominion makes the cat experience some sensations. The possession
involved here is thus abstract and the reference point is conceptualized as
experientially active.  Indirect object constructions are commonly used to
express some of these experiencer based relationships, e.g. Me duelen los 
ojos `My eyes hurt'. That the indirect object construction involves some 
abstract possession can be attested by the fact that the clitic me takes over 
the possessive marker's functions: 

5.a.Duelen mis ojos > Me duelen los ojos 
hurt 3rd-pl poss 1st-pl eyes   io-1st hurt 3rd-pl the-pl eyes
`My eyes hurt'
5.b.Corte mi       dedo>Me     corte el  dedo
cut 3rd-sgposs-1st-sg fingermddl-1st cut-1st    the finger
`I cut my finger'

As is well known, in standard dialects of Spanish, the indirect object 
precludes the use of the possessive marker: *Me duelen mis ojos. It is not the
case that the indirect object construction is in itself possessive, but that thetype of activity of the experiencer participant in indirect object constructionsimplies an abstract possessive relationship which entails that: a) the direct 
object is located with respect to its reference point; b) it is located in the
indirect object's dominion; and c) the experiencer/recipient establishes
contact with the direct object, receiving a concrete or abstract element and
undergoing some mental, judgmental or emotional experience (tell, send, give,
etc.) as a result of establishing contact with an element located in its
dominion. For example in: 
6. Arturo le dio un regalo a Blanca
`Arturo gave a gift to Blanca'
7.Juan Carlos por fin le dijo la verdad a Dora
`Juan Carlos finally told the truth to Dora'

it is clear that the gift changes from the possession of Arturo to that of
Blanca: the gift is first in the sphere of control--the dominion--of the
subject and ends in that of the indirect object. In a more abstract manner, the truth is located in the dominion of Dora--although not removed from Juan Carlos'--such that she now establishes mental contact with it. It is due to the fact 
that the direct object is located in the experiencer's dominion that s/he may 
interact with it. In Le envie una carta a Jorge `I sent a letter to Jorge' the
letter lands in an area where Jorge has some command and exerts dominion over
the object located in his sphere of action/control. Jorge is presumably active 
not only in keeping the letter in the goal of destination but also in reading 
and processing the information conveyed in it. The Reference Point Model 
involves then two levels of abstraction. From the conceptualizer's viewpoint
the reference point is a salient entity which is accessed in order to identify
an object located in her/his dominion. From the objective arrangement of the
event, the reference point is prototypically an active participant--thus its
high level of salience--exerting control over an object located in her/his
sphere of control (her/his dominion). Both characterizations interact in the 
construction of abstract possessive relationships.  The connection between levelof activity of a participant and dominion is an important one. An experiencer 
can only be active if an interacting element is within his/her dominion. In lackof a delimiting dominion the level of activity of the experiencer is limited. 
Consider the following examples:

8.a.Le di un regalo al presidente
`I gave a gift to the president'
b.Le abri la puerta al presidente
`I opened the president the door'
c.Abri la puerta para el presidente
`I opened the door for the president'
d.El ruido molesto al presidente
`The noise bothered the president'

The level of activity of presidente is not the same in all the examples. In 
(8.a) presidente is a prototypical indirect object, which actively holds
dominion over the direct object. Presidente is the beneficiary of the subject's action in both (8.b and 8.c); they differ however in level of activity. While 
in (8.b) it is implied that the president was about to walk through the door,
in (8.c) such implication is not present. The notion of benefaction is stronger when the action directly enters the dominion of the experiencer who is
therefore an active recipient in the development of the event. In (8.c) only
the subject's intentions are profiled by para the question as to whether the 
president will receive the subject's action remains open. In the cases
where le is used, the coexistence of the experiencer and the object in the same dominion allows the former to interact with the latter. Such coexistence is not evoked by para. The consequence of this is that the level of experiencer/object interaction is higher when le is used. A handful examples can be given with
the same type of contrast: Le lei un libro a los ninos `I read the children a
book'/ Lei un libro para los ninos `I read a book for the children' where only 
in the first case can the children be characterized not only as beneficiaries
but as good listeners. In lack of le only the subject's activity is 
foregrounded, while the potential activity of the beneficiary remains in the 
background. Finally (8.d) is the extreme case where the experiencer is a 
passive element in object position, as he simply undergoes a change imposed by 
an external impulse. It is only when the object of benefaction is located in 
the dominion of the experiencer that this participant can be seen as actively
involved with the object. The contrast between active and passive participants 
is represented in the following diagrams:







I will consistently represent the dominion in a dotted oval. Interaction is
represented with a dashed arrow.  As represented in Figure 3, participants are
active, as in (8.a and 8.b), depending on the fact that an object is located in their dominion. In the case of passive participants either the object is out 
the participant's sphere of control as in (8.c) or simply no dominion notion 
can be appealed to, as is the case of (8.d) and of subjects of passive
constructions.  In section III, I will show that the use of the middle marker 
se increases the level of activity of the experiencer participant; depending on the meaning of the verb, this phenomenon has direct consequences with respect tomood choice in Spanish. Before doing so it is necessary to establish a basic
characterization of mood in Spanish. In the following section, I show that the
indicative/subjunctive contrast is naturally accommodated by the notion of 
dominion.


II. Mood choice
It has been shown from different perspectives that mood choice is dependent uponthe level of assertiveness of the proposition as a whole. With respect to
Spanish, Terrell and Hooper pointed out that "the choice of mood in Spanish is
directly correlated with what the sentence as a whole expresses about the truth of the proposition included in the sentence"(1974:484). One may assert a 
proposition or presuppose the truth of a proposition in a number of ways: 
describing or commenting on a mental or a subjective reaction about something 
or simply pointing out the existence of something as a fact. Common mood choice situations are direct commands: di/diga la verdad `tell the truth'; sal/salga 
de aqui `get out of here', which are speech-act-situation dependent (high level of formality triggers subjunctive), and biclausal sentences in which the level 
of assertion of the main clause with respect to the contents of the complement 
clause is at issue. I will leave aside the speech-act situation problem and 
concentrate on biclausal phenomena. Consider the following examples:

9.Es evidente que Susana solo quiere trabajar (ind)
`It is evident that Susana only wants to work'
10.Los politicos no toman en consideracion que a la gente no le interesa 
   votar (ind)
`Politicians don't take into consideration that people are not interested
in voting'

Common to these examples is that the complement clause is given as an
unquestionable existing fact. On the other hand, complement clauses marked with subjunctive give special prominence to events whose existence is questionable:

11.No creo que Susana quiera trabajar (subj)
`I don't think that Susana wants to work'
12.Los politicos quieren que la gente tenga interes en votar (subj)
`Politicians want people to take interest in voting'

Terrell and Hooper (1974) and Terrell (1975) account for these examples by
claiming that in indicative clauses the truth of a proposition is asserted
while it is not in clauses marked for subjunctive. Although to some extent the
analysis holds, it has well-known shortcomings: while mental acts pertaining to a presupposed proposition trigger indicative, emotional reactions and comments
take subjunctive:

13.Me he dado cuenta de que miento, siempre he mentido (ind) (song S. Rodriguez)
`I have realized that I lie, I have always lied'
14.Me molesta que mientas tan cinicamente (subj)
`It bothers me that you lie so cynically'

In both cases the content of the complement clause is presupposed, however they
behave differently with respect to mood choice. Acknowledging this limitation,
Terrell and Hooper claim that mental-act verbs seem more similar to assertions 
than comment verbs do. However the assertion/non-assertion contrast offers no
explanation as to why some verbs would seem more assertive than others and why
subjunctive would be the mood chosen for only a subclass of presupposed 
complements. I do not intend to list all the shortcomings of that theory, nor
do I want to offer a total explanation of the Spanish mood system. I will
limit myself to a basic suggestion that captures in general terms the 
indicative/subjunctive contrast.  Since the assertion/non-assertion contrast is an insufficient factor for mood choice in Spanish, a more abstract notion is 
necessary. I propose that the notion of dominion captures all uses of the
indicative/subjunctive contrast in a natural manner. The notion implies an 
abstract possessive relationship between a participant and some entity (be it a thing or a proposition). The notion of dominion is a crucial part of elaborated reality (ER) (Langacker in press, Achard in press). "ER is the set of
circumstances accepted by a given conceptualizer as being real. Elaborated 
reality is describable by a set of propositions, which correspond to the
meanings of finite clauses. Propositions therefore incorporate grounding
predications, i.e. such elements as tense, modality, and negation, which locate circumstances with respect to the speech event (the locus of immediate reality).Hence the circumstances comprising elaborated reality are not limited to events (and states) that have actually occurred, but also includes the notion that
certain conceivable events have not in fact occurred, that a future event has
some potential to occur, etc.  A finite clause describes some facet of
elaborated reality, and unless it is grounded, an event has no status (or
"address") vis-a-vis this domain." The notion of dominion refers to the
conceptualizer's capacity to actively control and manipulate a circumstance in 
order to assess its status with respect to elaborated reality. A clause must be grounded (finite) for this purpose, and finite verb inflection can be identifiedwith indicative (as opposed to subjunctive) mood. The claim can be made then 
that indicative mood signals a conceptualizer's dominion over the circumstance
described in a clause, implying both grounding and active control. This proposal
leads to a general characterization of the indicative/subjunctive contrast: 
indicative clauses are located within the crucial participant's dominion, while subjunctive clauses are outside the participant's dominion. The indicative/
subjunctive contrast is schematically represented in the following diagrams:
CC=complement clause 
C=conceptualizer. 












The circle with the double arrow indicates the crucial participant (generally
the subject) and her/his capacity to interact with other elements. The circle
with the squiggly arrow represents the complement clause. The contrast is ratherevident: while in the indicative cases the complement clause is seen as part
of the elements that the participant has command over, as they are located in
her/his dominion, in the subjunctive cases, the event in the complement clause 
is out of such dominion and his capacity to interact with them can only be done out of speculation.  The conceptualizer is the most immediate participant from 
whose viewpoint the content of the complement clause is evaluated. The global 
conceptualization of the sentence as a whole is obviously done by the speaker, 
yet the value of the complement clause is calculated from the viewpoint of a
specific participant for whom the complement clause is relevant. In the case of impersonal ser constructions, the most immediate evaluator is the speaker, as 
can be attested in the following examples:

15.a.Es evidente que la familia no quiere saber de ti (IND)
`It is evident that the family doesn't want to know about you'
16.b.Es posible que la familia no quiera saber de ti (SUBJ)
`It is possible that the family may not want to know about you'

In the case of personal constructions the crucial conceptualizer is a prominent participant most intimately related to the content of the event in the 
complement clause--in most cases the subject of the main clause.  The
participant's capacity to interact with the complement is determined by the 
meaning of the verb.  Assertion is then only one of the ways in which the
participant may interact with the complement clause.  Notice that this accounts for all the data seen so far. In (15.a) evidente allows for the complement 
clause to be located in the dominion of the conceptualizer; in (15.b) posible
does not; thus the indicative/subjunctive contrast. Furthermore (16) and (17) 
are naturally accommodated:

17.Graciela quiere que Abelardo la acompa$e (subj)
`Graciela wants Abelardo to accompany her'
18.Los pol!ticos piensan/creen que la gente es est#pida (ind)
`Politicians think/believe that people are stupid'

The meaning of the verb will determine whether the content of the complement
clause can be located within the participant's dominion: querer only gives 
information about Graciela's desires, however she has no control and no access 
to Abelardo's decisions. Whether the act comes true or not is totally out of
her command. In contrast pensar/creer designate the dominion over an event or 
state by a conceptualizer.  The relationship with the complement clause is said to be local with respect to the conceptualizer. A number of circumstances in 
elaborated reality may determine the truth of a proposition, yet the actual way
it is seen from the conceptualizer's viewpoint is what determines mood choice.  Notice that the notion of dominion accounts for the indicative/subjunctive
contrast observed in complements clauses of saber `know':

18.a.Ya sabia que Antonio era (*fuera) medico (ind)
`I already knew that Antonio was a physician'
b.No sabia que Antonio fuera medico (subj)

`I didn't know that Antonio was a physician'
c. %No sabia que Antonio era medico (ind)
`I didn't know that Antonio is a physician'

The affirmative sentence in (18.a) is acceptable only in the indicative. The
conceptualizer possesses the relevant information to consider the complement
clause as a fact. In negative sentences, the subjunctive is the default case 
while the indicative is more context dependent. Both facts are accounted for by the notion of dominion. In (18.b) the content of the complement clause is
outside the conceptualizer's dominion, such that the truth of Antonio's being a physician cannot be taken as a fact; thus the complement clause is marked for
subjunctive. Now (18.c) is the case where the content of the complement clause
is in the dominion of a conceptualizer other than the speaker. In the context ofa conversation (18.c) is perfectly grammatical. What is at play here is that
the speaker does not present himself as the crucial conceptualizer. Actually, 
he asserts his lack of knowledge with respect to some proposition already seen
as true by the hearer. The use of indicative then highlights the fact that the 
complement clause is in the hearer's dominion--not in that of the speaker. The 
negative form is imposed on the activity of a conceptualizer that is not the
crucial one for the purposes of mood choice. Thus the subjunctive form is
ruled out for that particular situation. In an analogous manner, the
presupposition cases mentioned before, and repeated here for convenience, 
are accounted for naturally:

19.Me he dado cuenta de que miento, siempre he mentido (ind) (song S. Rodriguez)
`I have realized that I lie, I have always lied'

20.a.Me molesta que mientas tan c!nicamente (subj)
`It bothers me that you lie so cynically'
b.Ojala que puedas venir (subj)
`I hope that you can come'
c.Es una tristeza que no puedas cenar con nosotros (subj)
`It is sad that you can't have dinner with us'

In the indicative example the participant has enough dominion over the 
situation to present the complement clause as a fact. In the subjunctive
examples the content of the complement clause is out of the participant's 
dominion: in (20.a) me is a direct thematic object affected by the clause. 
Given that the participant is a non-active undergoer he is not seen as capable
of establishing any kind of control over any other element. In the other two
examples, the speaker manifests his/her reaction with respect to the
possibility of an event coming true. While in (20.b) and (20.c) the content of
the complement clause is not in the dominion of the crucial conceptualizer and 
is therefore not available to be considered as a fact, in (19) the content of 
the complement clause precisely enters the participant's awareness such that 
de que miento is seen as an unquestionable fact; mood choice is thus naturally
explained. Schane (in press) proposes the notion of subject responsibility to 
account for the indicative/subjunctive contrast. The notion of dominion is 
compatible with Schane's proposal, yet it is made more abstract in order to
cover other cases--such as temporal sequencing--which may be problematic for
the subject responsibility hypothesis. The well known temporal phrases 
introduced by cuando and other adverbial markers bear out the dominion analysis.Suppose I say:

21.Estoy seguro de que ire/voy a ir (ind)
`I'm sure that I will go/going to go'

My actual "going" may or may not take place, yet its occurrence does not
determine the selection of indicative mood in the complement clause. Instead 
it is the dominion I hold over the elements of present reality that allows me
to see future events as predictable facts. In contrast with this situation,
observe the following case:

22.Ira(ind) a verlo cuando termine de trabajar (subj)
`She will go to see him when she finishes working'

Although the act of going constitutes a potential fact, its realization depends on the completion of work.  The speaker holds enough information regarding the
fact that the participant will go, but the speaker's knowledge of present 
reality does not allow him/her to tell when the conditioning factor will be
under the subject's control, such that the event can be seen as true. The
contrast with the indicative use illustrates this fact:

23.Va (ind) a verlo cuando termina de trabajar (ind)
`She goes to see him when she finishes working'

The subject's completion of work happens iteratively such that this act
constitutes part of the events that the conceptualizer sees as facts of his 
everyday reality: it belongs to the set of elements located in his
dominion. It is thus expected that the indicative form in the complement 
clause will combine with the habitual reading imposed by the present tense of
the main verb. Things and actions that we control and conclude are 
conceptualized as being in our dominion and are therefore marked for 
indicative, those outside our dominion receive subjunctive marking.
The use of subjunctive with relative clauses exemplifies this contrast clearly:

24.Tengo un coche que es lentisimo (ind)
`I have a car that is very slow'
25.Quiero un coche que sea mas rapido (subj)
`I want a car that is faster'

The qualification of the object as slow in the indicative can only take place 
due to the fact that the concept of coche is in the subject's dominion. 
Referring to an element extraneous to that dominion can only be conjectural, 
thus the subjunctive marking in (25). Guitart, in this volume, proposes that
mood choice in relative clauses does not depend on the relationship established with the noun anteceding the relative clause but to the whole unit composed of
the noun and its modifying clause--I use Guitart's abbreviation NPR (Noun PhraseRelative) to refer to the whole unit. In (25), for example, the relationship is not with an abstract coche but with a particular car that has the property of
running fast. In agreement with Guitart's observation, I propose that if the 
content of the whole NPR is in the subject's dominion the relative clause is 
marked for indicative; while subjunctive mood signals NPRs that remain outside 
the subject's dominion.  Guitart, in this volume, convincingly rejects Bull's 
(1965) experienced/non-experienced hypothesis and has proposed alternatively
that it is only when the "NPR  refers to one or more entities that are
individuated in the speaker's mind...that the clause will be in the indicative."(in press)  While a type cannot take indicative, a specific instance of that 
type can, as can be seen from the contrast between (25 and 26): 

26.Quiero el coche que es mas rapido (subj)
`I want the car that is faster'

and from the following contrast:

27.a. Busco un maestro que sepa espanol (subj)
`I'm looking for a teacher that [may] know Spanish'
b. Busco un maestro que sabe espanol (ind)
`I'm looking for a teacher that knows Spanish'

The level of individuation of an object constitutes an important part of a
conceptualization. A participant interacts better with objects properly defined in his dominion than with those whose level of individuation is low. This
phenomenon is also important for the correct understanding of middle se, 
whose fundamental function is to increase the level of transitivity of the
clause; i.e. the level of interaction of an active participant with respect to
an object. Hopper and Thompson (1980) have shown individuated objects allow
for a greater degree of transitivity: vi un bailar!n `I saw a dancer' > vi al 
bailarin `I saw the dancer'. In the following section I show that se imposes a 
transitivity increase and derives constructions in which the participant is
more active. I will propose that the me temo type of constructions constitute 
extensions from middle se with high level of transitivity and that this 
phenomenon correlates with the use of indicative mood.


III Transitivity Increase
In section I, I proposed that when the notion of dominion is called upon the
level of activity of the experiencer is high. I have just noted that there is a high level of transitivity when the object of interaction is clearly 
individuated in the experiencer's dominion. Indirect object constructions have
been characterized as having active experiencers. In the case of the reflexive 
indirect constructions the subject and the recipient indirect object are equatedwith the same participant: Juan le dio un regalo a Maria `Juan gave a present 
to Maria > Juan se dio un regalo `Juan gave himself a present'. That the 
subject is the beneficiary and recipient of the gift is marked by the pronoun 
se. There are other constructions in which self-benefaction or self-orientation of the action is inherent to the meaning of the verb: conseguir `get, obtain',
reservar `reserve' and ahorrar `save' are obvious examples:

28.a. Consiguio un trabajo a dos cuadras de su casa
`She got a job two blocks away from her place'
b.Ahorre dinero, compre en nuestra gran barata
`Save money, buy in our great sale'

Since in conseguir the recipient and the subject are the same individual, 
rather than marking coreferentiality, se strengthens a benefactive meaning 
already inherent in the meaning of the verb:

29.a. Se consiguio un trabajo a dos cuadras de su casa
`She got herself a job two blocks away from her place'
b. Ahorrese el esfuerzo de ir pagar a la oficina, mande su pago por correo
`Save yourself the effort of going to pay at the office, send your payment by
 mail'

The clitic se emphasizes the fact that the event constitutes a particularly 
beneficial act in favor of the subject. Since the times in which Bello suggestedthe term superfluo to refer to this function of se there has been a general 
tendency in the Hispanic linguistic tradition to consider that this middle
marker is meaningless. Such an assumption is of course misleading: not only 
does it miss important information about the emphatic import of se, but it also restricts the possibility of observing the wide range of constructions that 
derive from it. Only a few of those derived patterns will be pointed out in 
this section.  That se is not vacuous can be seen from the ungrammaticality 
of (30):

30.*Ahorre el esfuerzo de ir pagar a la oficina, mande su pago por correo
`Save the effort of going to pay at the office, send your payment by mail'

A more obvious example where se is clearly meaningful can be observed from the
reservar/reservarse contrast:

31.a. Reservamos una mesa para cuatro personas
`We reserved a table for four people'
b. Nos reservamos el derecho de admision
c.*Reservamos el derecho de admision
`We reserve ourselves the right of admittance'

The crucial fact is that se imposes a transitivity increase whereby not only
the subject interacts with the object, but also keeps the object away from any
other participant's reach. In Maldonado (1991) I explain the contrast between 
the unmarked and the middle marked use of reservar as a consequence of 
conceptual intimacy between the subject and the direct object: derecho is 
already the subject's belonging, while mesa is not. With reservar an object
extraneous to the participant is located in her/his sphere of control. With
reservarse the increase of control over an inherent belonging is taken to the
extreme of marking it as exclusive to a specific participant. An extension from this construction type is constituted by full exploitation middle constructions.In this case, the verb not only implies that the object is in (or brought into) the subject's dominion but also that the participant interacts with it in some
manner. Obvious examples of this are comer `eat', tomar `drink' and other verbs of ingestion, like fumar `smoke', devorar `devour', etc.:

32.Abelardo fuma paquetes al dia
`Abelardo smokes packs a day'
33.Platero no toma alcohol
`Platero doesn't drink alcohol'

Provided that the object constitutes a well-defined and delimited unit, the
middle marker se imposes a completive interpretation:

34.Abelardo se fuma dos cajetillas diarias
`Abelardo smokes two packs a day'
35.Platero acababa de beberse dos cubos de agua (J.R. Jiminez)
`Platero had just drunk up two buckets of water'

The subject maximally exploits a well-delimited object located in his/her
dominion. What se indicates is that the action is done to completion and 
consequently that the whole direct object is affected by the subject's action. 
This value of se also occurs with verbs of mental activity:

36.Tengo alli un primo que se lo conoce todo: teatros, cabarets... -Se sabe 
cada sitio! (From Cartagena, 1972)
`I have a cousin that knows it all: theaters, cabarets... he knows some places!'
37.Pilar se sabe la cancion
`Pilar knows the song by heart'

Examples (36) and (37) contrast with their unmarked counterparts Conocer todo
`know everything', Saber la cancion `know the song' in that  the control of the subject over the direct object is maximal only in the middle constructions. 
Partial or limited dominion over an object is expressed in plain transitive
constructions:

38.a. Conoce un poco la ciudad
b.*Se conoce un poco la ciudad
`He knows the city a bit'

Since se increases considerably the transitivity effects of the verb, as in
the case of relative clauses, the object of full exploitation constructions 
must constitute delimited units properly contained in the participant's
dominion. This can be seen from the following ungrammatical tokens:

39.*Abelardo se fuma cajetillas diarias
`Abelardo smokes packs a day'
40.*Platero acababa de beberse cubos de agua 
`Platero had just drunk up buckets of water'
41.*Mi primo se conoce lugares
`Mi cousin knows places'

42.a. Maria sabe un poco de frances
b. *Maria se sabe un poco de frances
`Maria knows a little French'

Since the verb already implies some interaction of the subject with an object
located in his dominion (see examples 32 and 33), what se does is to give 
special prominence to the subject's action such that the object's attributes 
are maximally exploited (footnote 4).  The point being made here is that,
instead of being superfluous, se increases the level of interaction between
subject and object.  When the object constitutes a well-defined unit in the
participant's dominion the transitivity increase leads to a full exploitation 
reading. The facts seen so far suggest that such increase is a gradual one: 

plain transitive
reservamos una mesa / conoce un poco la ciudad
emphatic benefactive
nos reservamos el derecho de admision
full exploitation
se sabe la cancion

We may now be in a position to explain why the use of se determines the use of
indicative mood in the complement clause. In the following section I propose
that temerse and darse cuenta constitute extensions from full-exploitation and
that this fact determines the choice of indicative mood in Spanish.

IV. Mood choice and middle se
Common to all the middle values of se seen so far is that the clitic derives a 
new verb giving special prominence to some meaning already present in the verb. Yet the derivative value of se follows consistent patterns: in all cases it 
increases the level of activity/involvement of a participant, such that it 
designates a special benefactive reading in inherently benefactive verbs, and 
imposes a full exploitation reading in verbs in which the subject interacts 
with objects in her/his dominion. As will be seen, full-exploitation se
has applied to temer to derive temerse and this new verb prototypically 
determines the presence of an indicative complement clause:

43.a.Temo que las autoridades tomen represalias en contra de los estudiantes
 (subj)
`I fear that the authorities may take revenge against the students'
b. Me temo que las autoridades van a tomar/tomar n represalias en contra de los
 estudiantes (ind)
`I'm afraid that the authorities will/are going to take revenge against the 
students'

In absence of the middle marker me, the complement clause of (43.a) takes 
subjunctive, while in (43.b) the middle construction takes an indicative
complement. The translation is misleading in that it does not reflect the 
actual contrast between temer and temerse: temo expresses fear with respect to
some potential happening, while me temo predicts the happening of a future 
event. The middle use is a manifestation of full-exploitation se where the main clause subject holds enough information and participates in considering the 
event to a great enough extent that s/he is able to predict the future
existence of some act.  There is a coherence pattern in the development of this construal coming from two different directions: on the one hand, 
full-exploitation se imposes a reading of complete command over an element 
located in the subject's dominion; on the other, indicative mood choice
constitutes a level of assertion that depends on the location of an event withinthe speaker/participant's dominion. Since full-exploitation se implies
enough information about elaborated reality that a future event is predicted
and even seen as a fact, the choice of indicative mood in the complement clause is accounted for. Crucially, in (43.a) the notion of dominion cannot be 
appealed to since no middle marker is present. The subjunctive use is naturally
accounted for in absence of any delimited object located in a specified 
dominion. A more difficult question to be addressed is whether the combination 
of middle marker and subjunctive mood will lead to ungrammatical results. The 
answer is that it does not. This can be seen in the following example:

44.Me temo que las autoridades vayan a tomar/tomen represalias en contra de
 los estudiantes (subj)
`I fear that the authorities may take revenge against the students'

In fact Bello (1875), using a similar example (Me temo que os enganeis (subj)
`I fear that you may fool yourselves'), was the first to point out the existenceof what he called a superfluous dative marker whose value he described in the 
following manner:

    "con el se indica el interes de la persona que habla en el hecho de que se
     trata [with it one indicates the interest of the person who speaks about 
     the event at issue]" (1875:218).

Bello's description could not be more precise. Notice that the meaning of (44)
is closely related to (43.a), but quite distant from (43.b). The dative me of
(44) designates an increase of interest of the participant, much like it 
intensifies self-benefaction in consegui un trabajo `I got a job' > me consegui un trabajo `I got myself a job'. Two values of se seem to be present: a) an 
emphatic benefactive se and a full exploitation se. These values follow the samegradual organization exposed in section III. With respect to temer three levels of participation can be seen. In the plain transitive construction the
participant's level of involvement is low. As a benefactive marker, se 
increases the level of involvement of the participant.  A higher level of
participant involvement is developed in full-exploitation se. In temo the 
subject fears some remote possibility for an event to come true. The level of
involvement increases as me is introduced in (44). The participant's fear with 
respect to the potential happening of an event is greater, yet the
possibility for it not to happen is still open. While the event in the 
complement clause is equally remote, the subject's interests play a more
prominent role:

45.No he podido dormir utlimamente, me temo que un ladron vaya a venir en la
 noche
`I have not been able to sleep, I [myself] am afraid that a thief may come 
during at night'

where me emphasizes the speaker's concerns, as recurrent or rather intense.
Finally the predictive reading by which a future event is taken as an almost 
unquestionable fact emerges when full-involvement se is at play. It is the 
participant's control of facts in his/her dominion that allows his/her
prediction of a future happening. Under those circumstances, indicative mood 
in the complement clause is required. This reading is not available for temer 
without se:

46.*Temo que va a venir (ind)
`I fear that he is going to come' 

That the participant's level of control over the elements in his sphere of 
control is maximal in the middle/indicative combination construction, and that
it leads to a predictive reading, can be seen from juxtaposed expressions like
Es ya casi un hecho `It is almost already a fact' which assert the truth of the
preceding sentence:

47.a. Me temo que vas a reprobar/suspender matematicas. Eso es ya casi un hecho.  (ind)
`I am afraid that you are going to flunk math. That is almost already a fact'
b.*Me temo que vayas a reprobar matematicas. Eso es ya casi un hecho. (subj)
`I am afraid that you may flunk math. That is almost already a fact'

These facts suggest an organization of the participant's increasing level of
activity/involvement in the following manner:

unmarked/subjunctive > low involvement
Temo que venga
`I fear that he may come'

benefactive/subjunctive > higher involvement
Me temo que venga
`I fear [myself] that he may come'

middle/indicative > full involvement
Me temo que va a venir
`I'm afraid that he is going to come'

References to the predictive meaning are not lacking in the relevant 
literature. Molina Redondo (1974) has already pointed out that: "Temer tiene
dos significados: `tener miedo ' y `sospechar que pueda producirse algo desfavorable, no deseado, etc.'; temerse solo el segundo' [Temer has two meanings `have
fear' and `suspect that something disfavorable, unwanted, etc may happen';
temerse only has the second meaning]". It is clear then that the middle se 
marker imposes a special reading not available in the unmarked form of the
verb (footnote 5). Molina Redondo (1974) treats this phenomenon as a lexical
contrast; other authors (Alcina & Blecua (1975), Academia (1982) suggest that 
in temerse there is an important link between the so called "reflexive marker' 
and the use of indicative, however no particular explanation has been offered. 
Such a link is naturally accommodated as an extension that develops from the 
full-exploitation middle construction the way I have proposed here.  The 
contrast proposed in (47.b) is a gradual one. This means that the borderline
cases are harder to identify than those that are maximally opposed. This is 
precisely so. Speakers of Spanish have no problem in recognizing the two polar
extremes: the unmarked/subjunctive construction involving low involvement,
on the one hand, and the middle/indicative construction giving prominence to
the maximal involvement predictive reading, on the other. However the 
recognition of the intermediate stage of examples like me temo que venga (subj) not only takes more time to be processed, but also a special effort is needed to
identify its meaning.  Two polar phenomena can now be explained. First, the
fact that darse cuenta determines indicative mood choice as in (48.a), and 
second, that alegrarse, entristecerse and the whole class of verbs of emotional
reaction favor subjunctive complement clauses, as in (4.a), repeated here for
convenience as (49.a): 

48.a. Su madre se da cuenta de lo que quieren hacer (ind)
`Their mother realizes what they want to do'
b.*Su madre se da cuenta de lo que quieran hacer (subj)
`Their mother realizes what they may want to do'

49.a. Me alegro de que vengas (SUBJ)
b.??Me alegro de que vienes (IND)
`I'm glad that you are coming'

In both cases middle se derives a new verb from an unmarked transitive form. 
I will show that in both cases the level of involvement of the participant 
increases via the middle marker. Yet the difference depends on the semantic 
structure of the verb to which middle se applies. In the case of darse cuenta,
middle se applies to an already active participant, whereas in alegrarse it 
operates on a passive one:

50.Su madre da cuenta de lo que quieren hacer (ind)
`Their mother reports [gives account of] on what they want to do'
51.Me alegra que vengas (subj)
DO-1st-sing happy that come-2nd-sing
`That you are coming makes me happy [gladdens me]'

In (50) the mother has enough knowledge about her children to objectively
observe and report about their actions. The indicative mood is accounted for by the fact that the content of the complement clause is part of the elements the
mother has dominion over. Since se increases the level of interaction of the
subject with respect to the object, it is only natural that the derived verb
darse cuenta `realize' imposes a reading of awareness with respect to a mental
object. The subject's relationship with the object is more intimate in the
middle construction, as s/he assumes an experiencer role, than it is with the 
plain transitive verb, where the subject remains somewhat distant and is 
limited to observing and reporting about someone else's actions. Be it in the
plain transitive or in the middle construction, the content of the 
complement clause is located in the subject's dominion. The use of indicative
is thus accounted for.  The notion of dominion includes a reference point from 
which another object can be located. In a more concrete manner, the dominion 
constitutes the sphere of control/action where some participant is able to
interact with some object. The clitic me in (51) does not qualify as a referencepoint. It is a direct object, i.e. a passive participant undergoing an emotionalchange imposed by an external impulse. As a passive member no dominion and no 
sphere of control can be evoked. The subjunctive mood of the sentential
subject is accounted for by the fact that the content of the clause cannot be 
located in any participant's dominion. Now the function of middle se for this
type of verb is to transform a passive participant into a more active one. This fact can be seen from the following contrasting examples:

52.a. A Leon lo asusta que grites
`It frightens Leon that you scream'
b. Leon se asusta de que grites
`Leon gets frightened that you scream'

That Leon is more involved in experiencing a change of mental state in (52.b)
than he is in (52.a) has widely been acknowledged in the relevant literature
(Garcia 1975, Real Academia 1982, Alcina & Blecua 1975, Maldonado 1991). Leon
becomes more active as he changes from a passive undergoer to an experiencer 
participant. Yet the level of activity does not surpass that of a physical or
emotional reaction. (footnote 6)  Consequently the subject's level of activity
is not sufficient to evoke the notion of dominion, where s/he would interact 
with some object. His/her reaction is still impinged by an external cause 
which s/he may not control. The lack of interaction within a specific dominion
accounts for the use of the subjunctive mood of the oblique clause in (51)
and (52.b). All the examples seen here prove that level of activity of the
participant and the notion of dominion coordinate in determining the mood
choice in Spanish.  The behavior of middle se is then quite systematic. In all
the examples seen in this paper it has always been the case that it increases 
the level of activity of the most prominent participant in the main clause. 
Whether the activity increase leads to the use of indicative/subjunctive mood
in the complement clause will depend on the level of activity that the 
participant has in a specific construction before se applies to it. If the 
participant is already active, as in dar cuenta `account for', the activity 
increase imposed by se will determine indicative mood choice in the complement 
clause. If the participant is a passive affected entity, as in asustar,
alegrar, the middle marker se will increase her/his level of activity but not
to the extent of letting the participant be in control of the situation;
consequently, this type of middle construction will not determine the use of
indicative in the complementary oblique clause. Intermediate cases like temer 
do not fluctuate randomly. There are two ways in which the participant may 
increase his/her level of activity: a) augmenting his/her concerns about a 
potential event which remains out of his/her sphere of control and b) 
establishing control over a specific situation to the extent of being able to 
predict a future happening. It is only in the second case that a 
full-exploitation middle construction is at play. It is only then that the 
level of interaction of an active participant with a (mental) object located in his/her dominion will require an indicative complement clause. In explaining 
mental-act verbs (Se da cuenta de que + ind `S/he realizes that') versus 
comment verbs (Me molesta que + subj `It bothers me that'), Terrell and 
Hooper (1974) and Terrell (1975) suggested that mental-act verbs seem more
similar to assertions than comment verbs do. This paper explains Terrell and 
Hooper's intuitions and rounds out their findings with respect to the Spanish 
mood system. The similarity of mental-act verbs to assertions depends basically on the level of activity of the subject: with respect to an already active 
participant, full exploitation se derives highly active participants in control of their mental objects. This fact determines indicative mood choice. On the 
other hand "comment verbs" are far from being assertive precisely because the 
human participants of such verbs are conceptualized as passive.  The activating 
properties of se are insufficient to bring about a full fledged active 
participant and consequently no dominion can be called upon. The notion of 
dominion, as involving an active participant interacting with an object 
located in her/his sphere of control, has been offered here as a principled way to capture a set of intricate relationships between the middle marker se and 
mood choice in Spanish. It is my hope that the contents of this paper will 
shed some light on an area that was not completely covered by Terrell and
Hooper's proposal. 

Footnotes


1. I am indebted to Ronald Langacker for his guidance in this paper. I am also 
   in debt to Michel Achard, Aintzane Doiz-Bienzobas, Errapel Mejias-Bikandi, 
   Sanford Schane, Maura Velazquez for their insightful comments and to Donna
   and Valeria for breaking the rules I propose in this paper in their everyday 
   speech.

2. Clear cases in which se has a reflexive value are those in which there is a 
   split representation of the participant:

        1.      Era tal su soledad que Alcira se enviaba cartas a si
                misma (Ramirez Heredia El Rayo Macoy))
                `Her loneliness was such that Alcira would send letters
                to herself' 
        2.      Fue frente al espejo que Justine se dijo: ya me tienes
                harta, judia estupida (Durrell Justine)
                `It was in front of the mirror that Justine said to
                herself: you have me all fed up, you stupid jew'

   Most uses in Spanish and other Romance languages do not
   correspond to reflexive split representations but rather to a
   set of middle meanings in which some internally complex
   activity is being represented.

3. The indicative mood may be used for emphatic purposes either
   to signal that the subject of the complement clause is
   actually coming at that specific moment or to point out that
   his coming is an unquestionable fact.

4. Other details and extensions of this construction are given
   in Maldonado (1991), where a fine-grained study of full
   exploitation is presented.

5. However the possibility of using se with subjunctive is
   still available. In fact this variation suggests the origins
   of a quite recent grammaticalization process that has taken
   place in current Spanish. Although in-depth historical
   research on this problem is needed, it is possible that in the
   XIX century full involvement se had not yet extended to temer
   and that this derivation has been gaining ground in the course
   of the last century. 

6. No volitional or controlled action is executed by Le"n--the
   way it would be in constructions with active subject
   participants of the type Leon se seca la cara con una toalla
   'Leon dried his face with a towel'.


References

Achard, Michel. 1991. `Complement Construal in French and Spanish: A Cognitive
	Grammar approach' Papers in Honor of Tracy Terrell. New York: 
	McGraw-Hill.

Alcina, Juan and Jose Blecua. 1975. Gramatica espanola. Barcelona: Ariel.

Bello, Andres & Rufino Cuervo. 1845, 1951. Gramatica de la lengua castellana.
	Buenos Aires: 6th. ed. from the 1850 original. Caracas: Ministerio de 
	Educacion.

Bull, William. 1965. Spanish for Teachers: Applied Linguistics for Teachers of
	Spanish. New York: Ronald Press.

Contreras, Lidia. 1966. `Significados y funciones del se'. Zeitschrift fur 
	romanische Philologie. Tubingen: XC.298-307.

Garcia, Erica. 1975. The Role of Theory in Linguistic Analysis. Amsterdam and 
	Oxford: North-Holland.

Guitart, Jorge. 1991. `The NP-based Analysis of Mood Choice in Spanish
 	Relative Cluases'. Papers in Honor of Tracy Terrell. New York: 
	McGraw-Hill.

Hopper, Paul and Sandra Thompson. 1980. `Transitivity in Grammar and 
	Discourse'. Language 56-2.251-299.

Langacker, Ronald. 1987a. Foundations of Cognitive Grammar. vol. 1.
	Theoretical Prerequisites. Stanford: Stanford University Press.

-------1987b. `Nouns and Verbs'. Language 63.53-94.

-------1990. `Subjectification'. Cognitive Linguistics 1.5-38.

-------   in press. Foundations of Cognitive Grammar. vol 2: Descriptive
	Application. Stanford: Stanford University Press.

Maldonado, Ricardo. 1988. `Energetic Reflexives in Spanish'. Berkeley 
	Linguistics Society 14.153-165.

-------   1989. `Se gramaticalizo": A Diachronic Account of Energetic
	Reflexives in Spanish'. Proceedings of the IV Pacific Linguistics
	Conference, Eugene.  

-------   1991. Middle Voice: The case of Spanish se. Doctoral Dissertation. 
	San Diego: University of California at San Diego.

Mejias-Bicandi, Errapel. 1991. Papers in Honor of Tracy Terrell. New York:
	McGraw-Hill. 

Real Academia Espanola. 1978. Esbozo de una nueva gramatica de la lengua 
	espanola. Madrid: Espasa Calpe.

Schane, Sanford. 1991. `Illocutionary Verbs. Subject Responsibility and
	Presupposition: The Indicative vs.  the Subjunctive in Spanish'.
	Papers in Honor of Tracy Terrell. New York: McGraw-Hill.

Seiler, Hansjakob. 1983. `Possessivity. Subject and Object'. Studies in
	Language 7-1:89-117.

Terrell, Tracy. 1976. `Assertion and Presupposition in Spanish Complements'.
	Current Studies in Romance linguistics. (M. Lujan and F. Hensey. eds.).

Terrell, Tracy and Joan Hooper. 1974. `A Semantically Based Analysis of Mood 
	in Spanish'. Hispania 57.484-494.