Information structure

In addition to their syntactic structure, sentences also have an information structure, which concerns how the information they contain is related to the knowledge states of the participants in the discourse in which the sentences are used. These knowledge states in turn depend on the previous discourse. From an information structure perspective, sentences can be divided in a first cut into focus and ground. The ground in turn can be divided into a link and a tail. The discussion of these concepts that follows is heavily indebted to Vallduví 1990.

Focus versus ground

Ordinary sentences can be divided into two parts: a part that contains background information that is presupposed, the ground, and a part that is intended to be particularly informative, the focus. A simple way to understand the focus-ground partition is to consider questions and the corresponding short and full answers. In a question, the focus is the unknown information expressed by the wh- phrase. The remainder of the question is the ground. A short answer to a question consists only of a focus. Repeating the ground of the question yields a full answer. In the following examples, the focus and ground are highlighted in red and blue, respectively.

() a. Question:   What animals detest the smell of citrus fruits?
b. Short answer:   Ordinary cats.
c. Full answer: Ordinary cats detest the smell of citrus fruits.

There is no one-to-one correspondence between syntactic structure and information structure. Rather, one and the same syntactic structure can be associated with more than one information structure, as is evident from comparing () and (). Even though (c) has the same syntactic structure as (c), its association with the question in (a) imposes a different focus-ground partition on it.

() a. Question:   What do ordinary cats detest?
b. Short answer:   The smell of citrus fruits.
c. Full answer:   Ordinary cats detest the smell of citrus fruits.

Sentences can consist of a focus only; the ground can be empty. In such cases, there is no distinction between a short answer and a full answer. An example is given in ().

() a. Question:   What happened?
b. Answer:   John called.

As usual, other focus-ground partitions for the same sentence are possible, as shown in () and ().

() a. Question:   Who called?
b. Short answer:   John.
c. Full answer:   John called.
() a. Question:   What did John do?
b. Short answer:   Call(ed).
c. Full answer:   John called.

In spoken language, the focus-ground partition is generally conveyed by intonation. In written language, where intonation is difficult to represent, it is possible to indicate a sentence's information structure by associating the focus and the ground with particular syntactic positions. For instance, English it clefts exhibit a syntactic frame consisting of it, a form of the copula to be, and the complementizer that. The complement of be is the focus, and the complement of that is the ground. Different focus-ground partitions give rise to distinct it clefts, making it possible to disambiguate syntactic structures that would otherwise be informationally ambiguous.

() a.   Ordinary cats detest the smell of citrus fruits. →   It is ordinary cats that detest the smell of citrus fruits.
b.   Ordinary cats detest the smell of citrus fruits. →   It is the smell of citrus fruits that ordinary cats detest.

It is worth noting that the syntactic form of it clefts imposes syntactic constraints that can prevent a particular focus-ground partition from having a corresponding it cleft. For instance, the complement of be must be a maximal projection and that requires its complement to be overt. The hypothetically possible it clefts in () are therefore ungrammatical.

() a.   They will buy the car (not steal it). → * It is buy that they will the car.
b.   John called. → * It is John called that.

Link versus tail

Effective information management hinges not just in storing information, but in making it accessible in a way that is appropriate for particular purposes or in particular contexts. This is why databases allow a user to sort the entries in the database by different properties. For instance, a marketing firm might want to target all potential customers by zip code for one client, but by age for another client. We can think of zip code and age as sorting keys that take scope over the remainder of the retrieved information, as in ().