9. PLAN DE GESTIÓN AMBIENTAL PARA EL PROCESO PRODUCTIVO DE LA
9.2 PLANEACIÓN ESTRATEGICA
9.2.1 PERFIL DE CAPACIDAD INTERNA (PCI)
In natural languages the following elements may be identified: first-order terms, second-order terms, third-order terms (adjuncts), and, of a higher order still, operators
(such as WH, Neg or focus). First-order terms correspond to first-order entities like discrete objects and individuals. Second-order terms correspond to second-order entities (states of affairs, events, processes, activities, relations, universal quantifiers and some adjuncts; in other words, predicates). Third and higher-order terms correspond to third or higher-order entities (concepts, propositions).
According to the CSD an element of a lower order cannot depend on an element of a higher order {hcfw cannot depend on WH, focus, Neg, etc.), only on elements of the same or of a lower order. Elements of a higher order can depend on operators of the same order or of a lower order (a UQ over a lexical predicate). Some examples are given in table 2.1
1 Type of term 2 Example
t Operators WH, focus, Neg
t Third-order term Predicates over sets of sets of individuals
Iterative adverbs, exactly 5 students, more than 5 students
t Second-order term Predicates over a set of individuals
Every, each, frequency adverbs, adjuncts, scopeless indefinites
t First-order term Lexical predicates i Boy X
Table 2.1
1 (t) = order relation (e.g. an operator is of a higher order than a predicate over a set of individuals).
2 (>L) = dependency relation (e.g. Neg can depend on an adjunct, but an adjunct cannot depend on Neg).
Now the theory is in place, let us come back to those structures which exhibited weak-island effects. I argue that the stranded indefinite in French WHl-in-situ questions introduces only a Skolem function, and does not introduce an existential quantifier (otherwise the island effects cannot be explained). WH phrases in situ denote fiinctions with contextually supplied domains and ranges. When a questioner is using a French WHl-in-situ question like (la), she asks the hearer to identify and then supply a function with the property that for a person as input, the function will return what that person is doing. The hearer has to supply the necessaiy information: not only an appropriate domain, but also linking the intended object to something that is already given. The domain may be provided by a suitable quantifier. If there is no quantifier present in the question, then the domain is provided by the context or alternatively via inferences.
The Skolem function introduced by the stranded indefinite is bound by the null WH operator while the argument of the Skolem function can be unselectively bound by an intervening quantifier provided that the quantifier element is of a lower or of the same order as the stranded indefinite. The stranded indefinite is a scopeless element, and thus behaves very much like an adjunct in that its scope is fixed locally.
The WH phrase, or, more precisely, the remnant, can depend on an intervening frequency adverb because scopeless indefinites and frequency adverbs are of the same order. In the following examples I use J for ‘depends on’ and for the element on which the stranded indefinite depends, i is the functional index showing that the functional variable is bound by the question operator, while j is the argumentai index showing that the argument variable is bound by the intervening scopal element. It may be assumed that movement of the operator leaves a complex trace behind or alternatively that Op simply binds the functional variable/:
i J
(72a) [cpO p i[ip Tu fais [vp TOUJOURS^ k-piSK f i ( x i ) quoil ...]]]]?
‘What do you always do (at weekends)?’
^ ‘Which is the Skolem function f such that you do f(x, things) always at weekends? ’ f: a Skolem function from people into the things they do,
i J
(73a) [cpO pi[ip Tu fais [vp SQUVENTj F v p FSK f ,(xj) quoil ...]]]]?
‘What do you often do (at weekends)?’
^ ‘Which is the Skolem function f such that you do f(x, things) often at weekends? ' f: a Skolem flmction from people into the things they do.
In (78a) and (79a), the stranded indefinite can depend on the UQ. This is because the stranded indefinite is of the same order as the UQ;
i J
(78a) [cpOpi[ip Tous LES ENFANTS^ ont fait [SK /i (xj) quoi]]]? ‘What did all the children do?’
(i) * ‘Which x ,x a thing, every y, y a child, didx. ’
(ii) ^ ‘ Which is the Skolem function f such that every y , y a child, d id f (x, things)? ’ f: a Skolem function from children into the things they did.
^ J
(79a) [cpOpi [ip Tu as donné TOysj,ES VÊTgMElffSj à [SK f i (xj) qui]]]?
‘To whom did you give all the clothes?’
(i) * ‘Which x ,x a person, you gave every y , y a piece o f clothing, to x. ’
(ii) ^ ‘Which is the Skolem function f such that you gave every y, y a piece o f clothing, to f (x, persons)? ’
f : a Skolem fimction from pieces o f clothing into the people they were given to.
On the other hand, the stranded indefinite cannot depend on negation, focus, iterative adverbs or quantifiers such as exactement cinq étudiants (‘exactly five students’) or plus de cinq étudiants (‘more than five students’). This is because, in those cases, the intervening element is of a higher order than the stranded indefinite. (2a), (23 a) to (23), (24), (76) and (77) are repeated here with relevant (partial) derivations:
i J
(2a) *[cpOpi[ip Tu ne fais PASj [SK/i (xj) quoi] ce soir]]]? ‘What are you not doing tonight?’
* ‘Which is the Skolem function f such that you don’t d o f (x, things) tonight? ’ f: a Skolem function fi^om people into the things they do.
To spell out the oddness of (2a): the WH phrase in situ introduces a Skolem function. This means that there is a presupposed domain and range. To negate that is a contradiction: it is like saying that there is an x, but also that there isn’t an x.
i J
(23a) *[cpOpi[ip A u c u n ÉTUDIANT; a lu [SK/i(x j) quoi]]]?
‘What did no student read?’
* ‘Which is the Skolem function f such that no student has re a d f (x, things)? ’ / ; a Skolem function from students to the things they ha\>e read.
Again the question in (23a) leads to a contradiction. A domain and a range are presupposed, and the questioner is negating the domain and the range, so the question does not make any sense.
: J
(24a) *[cpOpi[ip SeulementJEANf fait [SK /i(xj) quo!]]]?
‘What does only JEAN do?’
* ‘Which is the Skolem function f such that only JEAN does f (x, things)? ’ f: a Skolem function from people into the things they do.
4. J
(25a) *[cpOpi[ip MêmeJEANj fait [S K /i(xj) quo!]]]?
‘What does even JEAN do?’
* ‘Which is the Skolem function f such that even JEAN does f (x, things)? ’ f: a Skolem fimction from people into the things they do.
i J
(26a) *[cp0pi[ipll a BEAUCOUP, lu [SK/i(xj) quoi]]]?
‘What has he often read?’
* "Which is the Skolem fimction f, such that he often rea d f (x, things)? ’ J: a Skolem fimction from people into the things they read
i J
(27a) *[cpOpi[rp ft a lu [SK/i(xj) quoi]]]?
‘What has he seldom read?’
* "Which is the Skolem fimction f , such that he seldom rea d f (x, things)? ’ f: a Skolem fimction from people into the things they read
^ J
(28a) *[cpOpi[ipft a TROP; lu [S K /i (xj) quoi]]]?
‘What has he read too much?’
* ‘Which is the Skolem fimction f such that he r e a d f (x, things) too much? ’ f: a Skolem fimction from people into the things they read
; J
(76a) *[cp Opi[n> EXACTEMENT CINO ÉTUDiANTSj ont fait [SK / i (xj) quoi]]]?
‘What did exactly five students do?’
* ‘Which is the Skolem fim ction f such that exactly five students d id f (x, things)? ’ f: a Skolem fimction from students into the things they did
i J
(77a) *[cpOpi[ip Plus de cinq étudiants; ont fait [SK /i(xj) quoi]]]? ‘What did more than five students do?’
* ‘Which is the Skolem function f such that more than five students d id f (x, things)? ’ f: a Skolem function from students into the things they did
In summary, I have shown that the stranded indefinite in a French single-WH-in- situ construction can depend on some intervening scopal elements, but not on others. This follows fi'om the NSC and the CSD, the set of conditions on scope to which the antecedent-govemment condition can be reduced. The present analysis is therefore an attempt to reduce weak-island and intervention effects to scope.
Let us now compare the present proposal with Szabolcsi and Zwarts (1992-1993) on the one hand, and Honcoop (1998) on the other.