What do you think?
Current Studies in Linguistics
Situations and Individuals (Current Studies in Linguistics
In Situations and Individuals, Paul Elbourne argues that the natural language expressions that have been taken to refer to individuals--pronouns, proper names, and definite descriptions--have a common syntax and semantics, roughly that of definite descriptions as construed in the tradition of Frege. In the course of his argument, Elbourne shows that proper names have previously undetected donkey anaphoric readings.This is contrary to previous theorizing and, if true, would undermine what philosophers call the direct reference theory (which holds that the sole contribution of a proper name to the truth conditions of a sentence is an individual) as well as the related doctrine that proper names are rigid designators. Elbourne begins by addressing donkey anaphora, relating other concerns about pronouns to the solution of this notorious problem. His subsequent argumentation provides a unified semantics for the donkey anaphoric and bound and referential uses of pronouns and discusses the prospect of unifying the syntax and semantics of pronouns with the syntax and semantics of normal definite descriptions. Elbourne's aim is not only to advance his proposal of a unified syntax and semantics but also to urge linguists and philosophers dealing with pronoun interpretation to consider a wider range of theories than they do at present, and to test the competing claims of description-based theories and dynamic semantics against the data.
- GenresLinguistics
234 pages, Paperback
First published November 1, 2005
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 2 of 2 reviews
July 27, 2026
Donkey pronouns are covariation without c-command.
One interpretation: those pronouns are DDs (aka D-type pronouns). Elbourne uses situation semantics and in particular situation extensions to satisfy the definedness condition (or “to neutralize the unwelcome uniqueness presuppositions that this move [scil. the identification of the semantic value of pronouns with DDs] produces”, p. 56).
Elbourne identifies 3 problems with the D-type approach: 1) the problem of indistinguishable participants (If a bishop meets a bishop, he blesses him), 2) the problem of the formal link (why the asymmetry between: Every man who has a wife sits next to her at dinner and *Every married man sits next to her at dinner?), and 3) the problem of pronominal ambiguity (D-type theories make pronouns ambiguous between individual variables and DDs).
Another interpretation is: dynamic semantics/predicate logic (DPL). In DPL, the meaning of a sentence is the way in which it changes variable assignments: it can be described as a function from variable assignments to variable assignments. “The system is such that for a pronoun to be coreferential with a previous expression, or to covary on the basis of a previous expression, the same variable that is used to translate the pronoun must have been introduced by the previous expression into the set of assignments that result from processing it.” (p. 35).
Elbourne identifies 3 problems with DPL: 1) the problem of disjunctive antecedents. There is no way of treating him in: If Mary hasn’t seen John lately, or Ann misses Bill, she calls him as a variable, since no variable has been antecendently introduced, only definites (in fact, proper names), and “dynamic systems do not normally say that definite terms introduce a variable” (p. 47). Another way in which variables can fail to be supplied is, not when there are only definite antecedents, but no linguistic antecedents at all, which is the second problem: 2) the problem of deep anaphora. The third problem is 3) paycheck pronouns, which are pronouns of laziness (Geach 1962): they can be substituted by exact repetition of their antecedents without meaning change. In: John gave his paycheck to his mistress. Everyone else put it in the bank, everyone only manipulates their own paychecks. But that means that it introduces a new entity—in fact, many new entities—the paychecks of people other than John. Yet nothing in the antecedent introduces the machinery required for it to be treated as variable in a way that outputs such an assignment.
In Ch 2, Elbourne introduces another possible treatment of donkey anaphora: the NP-deletion theory. NPs routinely undergo PF deletion under identity with a local NP (cf. My shirt is the same as his). Using this approach, in: Every man who owns a donkey beats it, it means the donkey. Hence the claim “that D-type pronouns can quite generally be viewed as being definite articles whose complements are subject to NP-deletion.” Basically we are using the identity condition on NP deletion (qua ellipsis) to generate strict readings of donkey (and other) sentences.
Ch 3 makes the analysis bear on the Frege v Russell debate regarding DDs: is ‘the’ a uniqueness presupposition (Frege) or a quantifier (Russell)? Elbourne argues the former, using ideas from Karttunen about how presuppositions project to the whole sentence when embedded by attitude verbs (John stopped drinking John used to drink; Mary believes that John stopped drinking Mary believes that John used to drink). Uniqueness presuppositions run into trouble trying to account for Senator Thad Cochran, the Mississippi Republican, announced that..., where [the Mississippi Republican] carries no presupposition that there is only one such (or: carries a ‘no uniqueness reading’), but “is merely taken to remind us of Senator Cochran’s home state and party affiliation” (p. 155). To justify the uniqueness presupposition, we could narrow down the universe of discourse in a way that excludes other Mississippi Republicans (as not at issue). That does not seem to be possible however. In Thad Cochran, Mississippi's Republican Senator..., the universe of discourse is narrowed down to Thad Cochran, yet the genitival DD cannot deliver a no uniqueness reading.
In an important passage, Elbourne argues (i) that there are two theories of pronouns: one according to which they take two arguments, an index i and a normal NP, and another according to which they take only one, either an index or a normal NP, and (ii) that, on the first view, it is not the case that a regular NP must be provided or filled in the structure either by linguistic or nonlinguistic context. Elbourne claims there is a contrast between two utterances in a context where you are passed by by a young, twenty-something Caucasian skateboarder wearing a baseball cap. The first utterance is: ?Most look more depressed than that, and the other is He looks happy!. The argument is that, although the description under which the young man is referred to by most and he respectively is equally obscure (on the one hand: most what? twenty-year-olds? young men? skateboarders?; on the other: the young man? the skateboarder? the young man who just passed us by?), only the first utterance is weird and hard to interpret, while the second is perfectly fine. Elbourne writes: “I conclude that any theory according to which pronouns take two arguments, just like overt definite articles, will have to allow the second argument to be a kind of default item which is always available and does not need to be recovered by means of a linguistic antecedent or overwhelming contextual salience.” (p. 164, my emphasis). The important conclusion is that it is not necessary to recover a given description as the antecedent of a pronoun in order to compute the semantics of a sentence, and that this is specific to pronouns, compared to quantifiers. Additionally (this is my thought), we could think of the pronoun instead as an individual-tracker, i.e. a pointer, anchored to its denotation (the young, skateboarding, baseball-cap-wearing (etc.) man passing by in the example). The general conclusion is that nonlinguistic antecedents are not represented descriptively. In Elbourne's words, the nonlinguistic antecedents of pronouns are indices, not NPs, although they will be of type < e, t > (p. 167).
In Ch 4, Elbourne tackles the problem of indistinguishable participants, e.g. bishop sentences. Neale's (1990) descriptivist analysis runs into trouble accounting for plural vs singular forms. Heim (1990) proposes to quantify over minimal situation extensions: for If a man lives with another man, he shares the rent with him, what you do is quantify over situations such that for all minimal situations where one man lives with another man, there is an extension of that situation in which they both pay rent. The problem with this more sophisticated account is that it overgenerates uniqueness presuppositions. Indeed, the roommate sentence is not committed to there being no more than two roommates. Heim imagines a situation with three men living together, and two of them paying the rent, as compatible with the roommate sentence. In reply, Elbourne suggests that what a D-type theory of pronouns must account for in such cases is the acceptability of If a bishop meets a bishop, he blesses him, and the unacceptability of *If a bishop and a bishop meet, he blesses him, which, Elbourne suggests, is triggered by the fact that a definedness criterion or presupposition goes unmet. The solution Elbourne proposes turns on keeping the two bishops x and y in two, mutually isolated, minimal situations (s2, which contains x, on the one hand, and s4 and s5, which contain y and the meeting of x and y, but not x himself, on the other, in (33), from p. 193)
Elbourne illustrates his situation-theoretic way out of the problem of indistinguishable participants via an analogy with a sentence like: The table is wobbly. We should have kept Aunt Linda's oak table. The idea is that, in such a sentence, despite synonymity (two tables), there is a relative salience of the two tables that allows you to pick the right one out in a way that makes the truth conditions come out right. The idea is that, in bishop sentences, we have more of the same thing, even though both bishops are introduced right away. Still, Elbourne contends, “there is a basis for differentiating the two bishops, as already pointed out, and there is a some intuitive reason to suppose that what I have been calling the “distinguished” bishop in each case is more salient than the other one, since he is the odd one out, the one whose properties are partially barred from the large situations s5, where everything else goes on.” (p. 200). In plain terms, some partition of the logical space (into a hierarchy of situations) is going on, out of which participants emerge with distinct relative salience statuses, making individuation possible.
Elbourne's discussion here mentions Lewis (1973: 116), where a distinction is made regarding DDs between “uniqueness presupposition” and “presupposition of determinate uniqueness”, in which case “most reasonable resolutions [of reference] will yield the same unique most salient [F]”.
One interpretation: those pronouns are DDs (aka D-type pronouns). Elbourne uses situation semantics and in particular situation extensions to satisfy the definedness condition (or “to neutralize the unwelcome uniqueness presuppositions that this move [scil. the identification of the semantic value of pronouns with DDs] produces”, p. 56).
Elbourne identifies 3 problems with the D-type approach: 1) the problem of indistinguishable participants (If a bishop meets a bishop, he blesses him), 2) the problem of the formal link (why the asymmetry between: Every man who has a wife sits next to her at dinner and *Every married man sits next to her at dinner?), and 3) the problem of pronominal ambiguity (D-type theories make pronouns ambiguous between individual variables and DDs).
Another interpretation is: dynamic semantics/predicate logic (DPL). In DPL, the meaning of a sentence is the way in which it changes variable assignments: it can be described as a function from variable assignments to variable assignments. “The system is such that for a pronoun to be coreferential with a previous expression, or to covary on the basis of a previous expression, the same variable that is used to translate the pronoun must have been introduced by the previous expression into the set of assignments that result from processing it.” (p. 35).
Elbourne identifies 3 problems with DPL: 1) the problem of disjunctive antecedents. There is no way of treating him in: If Mary hasn’t seen John lately, or Ann misses Bill, she calls him as a variable, since no variable has been antecendently introduced, only definites (in fact, proper names), and “dynamic systems do not normally say that definite terms introduce a variable” (p. 47). Another way in which variables can fail to be supplied is, not when there are only definite antecedents, but no linguistic antecedents at all, which is the second problem: 2) the problem of deep anaphora. The third problem is 3) paycheck pronouns, which are pronouns of laziness (Geach 1962): they can be substituted by exact repetition of their antecedents without meaning change. In: John gave his paycheck to his mistress. Everyone else put it in the bank, everyone only manipulates their own paychecks. But that means that it introduces a new entity—in fact, many new entities—the paychecks of people other than John. Yet nothing in the antecedent introduces the machinery required for it to be treated as variable in a way that outputs such an assignment.
In Ch 2, Elbourne introduces another possible treatment of donkey anaphora: the NP-deletion theory. NPs routinely undergo PF deletion under identity with a local NP (cf. My shirt is the same as his). Using this approach, in: Every man who owns a donkey beats it, it means the donkey. Hence the claim “that D-type pronouns can quite generally be viewed as being definite articles whose complements are subject to NP-deletion.” Basically we are using the identity condition on NP deletion (qua ellipsis) to generate strict readings of donkey (and other) sentences.
Ch 3 makes the analysis bear on the Frege v Russell debate regarding DDs: is ‘the’ a uniqueness presupposition (Frege) or a quantifier (Russell)? Elbourne argues the former, using ideas from Karttunen about how presuppositions project to the whole sentence when embedded by attitude verbs (John stopped drinking John used to drink; Mary believes that John stopped drinking Mary believes that John used to drink). Uniqueness presuppositions run into trouble trying to account for Senator Thad Cochran, the Mississippi Republican, announced that..., where [the Mississippi Republican] carries no presupposition that there is only one such (or: carries a ‘no uniqueness reading’), but “is merely taken to remind us of Senator Cochran’s home state and party affiliation” (p. 155). To justify the uniqueness presupposition, we could narrow down the universe of discourse in a way that excludes other Mississippi Republicans (as not at issue). That does not seem to be possible however. In Thad Cochran, Mississippi's Republican Senator..., the universe of discourse is narrowed down to Thad Cochran, yet the genitival DD cannot deliver a no uniqueness reading.
In an important passage, Elbourne argues (i) that there are two theories of pronouns: one according to which they take two arguments, an index i and a normal NP, and another according to which they take only one, either an index or a normal NP, and (ii) that, on the first view, it is not the case that a regular NP must be provided or filled in the structure either by linguistic or nonlinguistic context. Elbourne claims there is a contrast between two utterances in a context where you are passed by by a young, twenty-something Caucasian skateboarder wearing a baseball cap. The first utterance is: ?Most look more depressed than that, and the other is He looks happy!. The argument is that, although the description under which the young man is referred to by most and he respectively is equally obscure (on the one hand: most what? twenty-year-olds? young men? skateboarders?; on the other: the young man? the skateboarder? the young man who just passed us by?), only the first utterance is weird and hard to interpret, while the second is perfectly fine. Elbourne writes: “I conclude that any theory according to which pronouns take two arguments, just like overt definite articles, will have to allow the second argument to be a kind of default item which is always available and does not need to be recovered by means of a linguistic antecedent or overwhelming contextual salience.” (p. 164, my emphasis). The important conclusion is that it is not necessary to recover a given description as the antecedent of a pronoun in order to compute the semantics of a sentence, and that this is specific to pronouns, compared to quantifiers. Additionally (this is my thought), we could think of the pronoun instead as an individual-tracker, i.e. a pointer, anchored to its denotation (the young, skateboarding, baseball-cap-wearing (etc.) man passing by in the example). The general conclusion is that nonlinguistic antecedents are not represented descriptively. In Elbourne's words, the nonlinguistic antecedents of pronouns are indices, not NPs, although they will be of type < e, t > (p. 167).
In Ch 4, Elbourne tackles the problem of indistinguishable participants, e.g. bishop sentences. Neale's (1990) descriptivist analysis runs into trouble accounting for plural vs singular forms. Heim (1990) proposes to quantify over minimal situation extensions: for If a man lives with another man, he shares the rent with him, what you do is quantify over situations such that for all minimal situations where one man lives with another man, there is an extension of that situation in which they both pay rent. The problem with this more sophisticated account is that it overgenerates uniqueness presuppositions. Indeed, the roommate sentence is not committed to there being no more than two roommates. Heim imagines a situation with three men living together, and two of them paying the rent, as compatible with the roommate sentence. In reply, Elbourne suggests that what a D-type theory of pronouns must account for in such cases is the acceptability of If a bishop meets a bishop, he blesses him, and the unacceptability of *If a bishop and a bishop meet, he blesses him, which, Elbourne suggests, is triggered by the fact that a definedness criterion or presupposition goes unmet. The solution Elbourne proposes turns on keeping the two bishops x and y in two, mutually isolated, minimal situations (s2, which contains x, on the one hand, and s4 and s5, which contain y and the meeting of x and y, but not x himself, on the other, in (33), from p. 193)
Elbourne illustrates his situation-theoretic way out of the problem of indistinguishable participants via an analogy with a sentence like: The table is wobbly. We should have kept Aunt Linda's oak table. The idea is that, in such a sentence, despite synonymity (two tables), there is a relative salience of the two tables that allows you to pick the right one out in a way that makes the truth conditions come out right. The idea is that, in bishop sentences, we have more of the same thing, even though both bishops are introduced right away. Still, Elbourne contends, “there is a basis for differentiating the two bishops, as already pointed out, and there is a some intuitive reason to suppose that what I have been calling the “distinguished” bishop in each case is more salient than the other one, since he is the odd one out, the one whose properties are partially barred from the large situations s5, where everything else goes on.” (p. 200). In plain terms, some partition of the logical space (into a hierarchy of situations) is going on, out of which participants emerge with distinct relative salience statuses, making individuation possible.
Elbourne's discussion here mentions Lewis (1973: 116), where a distinction is made regarding DDs between “uniqueness presupposition” and “presupposition of determinate uniqueness”, in which case “most reasonable resolutions [of reference] will yield the same unique most salient [F]”.
March 23, 2011
Elbourne often writes engagingly, which is a great help to what can be both a fascinating and dry subject. I'm afraid that I don't have much confidence in a lot of his data; there were times when I and several folk around me did not share (crucial) judgments. I'm not sold on the theory, either, but there are some wonderful insights, and I very much like tossing rigid designators out the window.
Displaying 1 - 2 of 2 reviews