Definition of Reference 48757 Stanley ref Manage in Olo

3 The Rationale for Different Referential Forms 3.1 Why Are There Different Referential Forms? Referential form and its management has in recent times been investigated in at least three separate dis- ciplines: artificial intelligence, psycholinguistics, and linguistics. The people in artificial intelligence want to understand the use of reference so that machines can parse or generate “natural” language. The human ability to use the same pronominal form to refer to many different objects, even in the same environment, is very difficult for AI researchers to duplicate with computers. Psycholinguists use language as a window into the mind. They try to account for the mental processes behind the use of language. Linguists fall into different camps. Some look at the study of reference as merely a tool to help analyze the structure of a lan- guage. Others look at reference as part of a complete description of a given language. And finally, some view reference in much the same way as psycholinguists, as a way to understand the human mind. While I have divided up the different reasons for looking at reference, this is not to suggest that a given researcher could not have multiple goals in his study of reference. For my own part I am inter- ested in reference as being part of a complete picture of a given language and as a window into the mind of language producers and comprehenders. I believe that the choice of referential form is cognitively based and conventionalized. It is cognitively based in that the use of a given referential de- vice has cognitive effects in the minds of the audience. Specifically the different referential forms cause increases in the activation level of their referents and cause suppression of the activation of all other participants. The amount of activation and suppression varies according to the specific referen- tial form used. Generally the larger and more specific the form, the greater the effects will be Gernsbacher 1990. Referential form is conventionalized in that there are rules or strategies 1 that govern the use of given referential devices. These rules or strategies need to be discovered for each given language. The strategies, while based on general cognitive principles, will vary from one lan- guage to another. For example, Perrin 1978 reports that the main character in Mambila folk tales is realized almost exclusively by a zero after the initial introduction. Other languages that have zero as a realization for a referent use it in different circumstances.

3.2 Definition of Reference

Before going on to examine the various theories of reference, it is necessary to be clear about just what a REFERENCE refers to. People often make the mistake that a particular instance of a referential 47 1 The distinction I am making between rules and strategies, is that a rule is automatic and a strategy is not. Rule governed behavior occurs when given x then y, and since y occurred x must have occurred. A strategy is probable so that given x then y is likely. It can also be the case with strategies that they are only one way conditionals, so while x then y is true it does not mean that finding y means that the occurrence of x is also true. form refers to a previous or subsequent instance of a particular referential form, ANAPHORA and CATAPHORA . This is clearly not the case. A reference is an expression which designates an object. It is in- correct to see it as something that points to another expression or as a definition of the object. Any dis- cussion about one word referring to a previous word, when in actual fact both are referring to something else, is at best misguided. It is not a new idea that reference is to some object. Since antiquity, philosophers have discussed the meaning of words and how words evoke the identity of the object being referred to. This can be partic- ularly problematical when a single word can refer to many different objects, or a single object can be referred to by many different expressions. In each case we would like to know how someone resolves the ambiguity. When a speaker uses a specific term to refer to an object, what influences his decision? How will his listener interpret what he said? These problems face us anytime we try to understand the reasons behind the choice. To understand the reasons we need to first understand what we are doing when we refer to some object. While people can refer to nonexistent objects like unicorns, let us set aside for the moment any dis- cussion of things that do not occur in the real world. Questions concerning objects and how people identify them have a long history. A common definition of a noun can be traced back to Aristotle in De Interpretatione On Interpretation McKeon 1941 involving a conventionalized sound combination that signifies something that is stable over time. In more recent times, other philosophers have under- taken to clarify the ideas about identity and how we identify. J. S. Mill 1843 proposed to distinguish two uses of nouns, what he called their “connotation” and “denotation.” In Mill’s use of the terms, a word like “man” denotes both the class itself as well as one of the members of that class. The word also “connotes” or implies those attributes of a man that allow an individual object to be recognized as an instance of the class. Russell 1905 built on Mill’s ideas of connotation and developed a theory that a reference to an object is equivalent to an adequate description of that object. He essentially builds his arguments around the truth value of a noun phrase as in 139. 139 The King of France is bald. The King of England is bald. Since there is no King in France, the first statement cannot be significant 2 according to Russell, since it cannot refer to anyone; while the second is significant because it does refer to someone at least in 1905 when Russell wrote. This is based around the belief that the logical form of the expression “The King of France” entails “There is a king of France.” Russell is advocating that an expression with a noun that denotes nothing must be false. He claims that a truth value is asserted for statements that de- note something. So while “the king of France” cannot have a truth value in itself, the statement it en- tails: “There is a king of France” clearly can. However as Aristotle pointed out in De Interpretatione On Interpretation McKeon 1941, nouns and verbs have in themselves no truth values. And further in De Sophisticis Elenchis On Sophistical Refutations McKeon 1941, Aristotle says the name of an item is distinct from that item, and what follows in the name of an item does not follow necessarily in the item itself. Contemporaries of Russell also took exception to his Theory of Descriptions. Strawson 1950 in particular points out these problems. He makes an important point about reference, saying, “‘Mention- ing’ or ‘referring’, is not something an expression does; it is something that someone can use an expres- sion to do” Strawson, 1950:326. Strawson goes on to point out that a particular utterance can only be interpreted inside of a context. He points out that an expression is geared to a particular setting, so that people can indicate and delineate the objects in that setting that they want to talk about. People use expressions to refer to particular things. But the meaning of an expression is not the set of things or the single thing it may correctly be used to refer to: the meaning is the set of rules, habits, conventions for its use in referring. Strawson 1950:328 Strawson acknowledged that people can and do talk about objects that do not exist as if they did exist, but he set that outside of the limits of his discussion. However, this needs very much to be inside the limits of the discussion of reference because any theory of reference must account for the behavior of language speakers. Kripke 1972 undertook the development of a theory of reference that would handle reference 48 The Rationale for Different Referential Forms 2 Strawson 1950 equates Russell’s significant with being meaningful. to non-existent objects. In doing this Kripke introduces the term RIGID DESIGNATOR . This term is used for names, which completely identify or specify the object to which the name has been given. This designation is given to the object, and it completely specifies that object. This holds for both contractual statements about some individual, like, “If Nixon had lost in 1968, he would have retired to California.” The name “Nixon” uniquely identifies the individual. And it is only by that unique identification that we can discuss what that individual would have done in some alternate world that is different in some aspect from the real world. If we could somehow change the essential character of Nixon so he would no longer be Nixon, but say Humphrey, this new person the Humphrey clone would no longer be Nixon, but someone else. This is why Kripke talks about names as rigid designators. This rigid designation also holds for completely ficti- tious characters like Sherlock Holmes. Holmes is the product of the mind of Arthur Conan Doyle. If some- one were to appear and claim that Doyle modeled Holmes after him in the most exacting detail, that would not make this person Holmes. Holmes is simply and completely the character envisioned by Doyle. If some- one is to claim that he is modeled after someone else, the term “modeling” itself is enough to see that the two are not the same, any more than a bust of Aristotle is Aristotle, or a statue of Paul Bunyan is Paul Bunyan. Not all nominal expressions are used to refer Donnellan 1972. Some of them, to use Donnellan’s term, are attributional. That is, when I say “John is a teacher,” I am not making a reference equating John and teacher, but I am describing John. It is important that we realize that not all nominals are used to refer. When a nominal is used attributively it is not necessarily able to be substituted for a ref- erential nominal without changing the meaning. This has been referred to as referential opacity Quine 1953. That there should be problems with such substitutions should not be surprising. Speak- ing involves communicating and therefore involves at least two people. No one speaks outside of a context, and it is the context that provides for the rigid designation. In the context of this philosophical discussion a mention of Alexander the Great would most likely hearken the reader back to an individ- ual who lived between 356 and 323 B.C., the son of Philip of Macedonia, and conqueror of “the world.” However, there is another Alexander the Great. On Internet a particular individual uses the nom de plume “Alexander the Great.” While he is not the Alexander the Great of antiquity, he is Alexander the Great of Internet and is commonly referred to as Alexander the Great. In this case the reference by a communicator only succeeds within a given context. Reference is in a sense negotiated by the speaker and hearer Clark and Marshall 1981. A person who wants to communicate will refer to someone in such a way that the listener will be able to identify the ob- ject of the reference. As Clark and Marshall point out, for a reference to succeed, the use of the referring ex- pression has to identify the same object for both sides in the communication exchange. Fauconnier 1985 makes the same point only in a broader context, claiming that the whole mental space must be negotiated for communication to succeed. One of the keys to this negotiation for Fauconnier is the identification of the referents based on the triggers provided by the speaker. Now an exchange can be about something that is present in the same physical location as the speech act participants, but it does not have to be. The better I know a particular individual, the better I can judge his knowledge of the world and refer to entities in a re- solvable way for my interlocutor. So for someone familiar with the University of Oregon I can talk about the “EMU” with no further specification. For someone familiar with universities in general I can talk about the student union, or the student union building. However, if I am discussing this with someone who is completely unfamiliar with university life in the United States, the term “student union” would not com- municate what I intended. Some items are universally known by mankind. Items like the sun or moon can be referred to without negotiation in that the assumption is made that the reference will always be explicit. Other items are more culturally bound. In Japan a reference to “the Emperor” would refer to a specific in- dividual. The same would hold true in the United States for “the president” or “Congress.” The use of these terms can be overridden by local context. So, I can refer to a specific individual who is not the President of the United States, by just using the term “the president” if the context has been properly set. Two univer- sity secretaries would have no trouble resolving the following utterance to refer to the university president. 140 The president requested all the deans to meet with him at 2:30. Reference is made to objects that are resolvable by the participants in a communication. The com- munication can be about real objects or imaginary objects. A reference designates an object. The 3.2 Definition of Reference 49 object may exist in the real world or only exist in the universe of the discourse Givón 1989. The uni- verse of discourse is a shadowy affair, what Givón calls “crypto pragmatic,” in that it is brought into being by the speaker. Many things are mapped from the real world into this universe of discourse, but other things may only exist inside this universe. Both are discussible because they are identified as co- herent “objects” within the context that the speaker and hearer have established, Fauconnier’s 1985 mental space. For purposes of the discourse these objects are treated the same as “real” objects. The participants in a speech act in a sense negotiate the objects in the world of discourse. When one wants to tell a fanciful story, he asks his listeners to suspend their disbelief Coleridge 1905 and treat the ob- jects and sometimes even the “natural” laws that the speaker invents as if they are real. Since we can have wholly fanciful objects in a story, we must also allow that when we refer to “real” objects we are at best pointing out something in the environment that can be perceived. We can also be marking out something that is not readily perceivable. This can be the result of it not being observable for one reason or another. The reasons that something is not observable can vary from it not being in the same spatial location during the speech act to it being some kind of abstract concept. Both the lan- guage producer and the language comprehender create mental representations of the participants and events, what Fauconnier 1985 calls mental space. It is these mental representations that are talked about. An ideal communication would allow the mental representation of the producer to be given without any changes or deletions to the comprehender. Obviously all communication fails to some ex- tent in accomplishing this. In all cases by referring to something we are proposing to our audience that they accept it as “real” for the purposes of the discourse we are engaged in. The universe of discourse is this mental space containing all the negotiated objects in it. In this way there is no need to distinguish “real” from imagined objects when studying discourse. A distinction must be drawn between this universe of discourse and the text, and further between the objects referred to and the textual references. The text is the verbal account of the communication. It is not the universe of discourse, although it is part of it. The universe is the context of the text, and so any part of the text which precedes the current utterance is part of that context. A common misconcep- tion deals with what a reference in a text “refers” to. This is often aided and abetted by the two terms of reference, anaphora and cataphora. The terms involve direction in the text. What should be clear at this point in the discussion is that a reference is used by a speaker to refer to a referent. That is, a speaker uses an expression to designate a particular object or group of objects. In this way, if a speaker uses two separate expressions to designate a single object, the reference of both is to the object, not from one to the other. The terms anaphora and cataphora are useful in discussing how those situations in which a particular expression, particularly a pronoun, is resolved to identify the referent it desig- nates. If the information needed to resolve the reference is in the previous discourse, the pronoun can be considered anaphoric, while if it is in the following discourse it is considered cataphoric. Thus, the distinction between an anaphoric reference versus a cataphoric reference is one of direction. In gen- eral, a better characterization of reference is that it is EXOPHORIC , that is, reference actually points out- side the text. Two references are coreferential if they both refer to the same participant, and not the exclusive province of one referring to the other. This distinction is important as it is needed to explain definite references which have no text antecedent, which Gernsbacher 1991 calls “conceptual anaphors.” An example of this is given in 141 Sidner 1981:217. 141 a. My neighbor has a monster Harley 1200. b. They are really huge but gas-efficient bikes. As Sidner points out 1981, 1984 the they in 141b is not an anaphor of a monster Harley 1200. The two differ in number so they are not coreferential. However, the use of the pronoun is not infelicitous. A native speaker readily understands the use of the pronoun to refer to the set of objects for which a monster Harley 1200 is an instance. If pronouns are anaphoric they must have an antecedent. It is clear from the example that there is no textual antecedent, and so pronouns are not necessarily anaphoric. Rather, they refer to something that is cognitively established in the mind of language producer and comprehender. The sentence in 141a establishes the Harley 1200 in the universe of discourse. It is 50 The Rationale for Different Referential Forms established as both an individual item, but also as a name of a class of items. The actual name is of the class not of the individual item in the class. It is natural then that the reference in 141b is made to the class with a pronoun as easily as it could be made to the item. From the preceding discussion we have defined the idea of reference to be used in this work. A refer- ence is an expression which designates an object. References are always made to objects in some uni- verse of discourse which may or may not be completely accounted for by objects in the real world. Two references are coreferential only if they designate the same entity, or two identical groups of entities. Any appeal to claims of one reference referring to another is clearly spurious. In this work I only use the terms anaphoric or cataphoric to mean those references which precede another coreference, or which follow a coreference.

3.3 Referential Choice and the Environment