For Bill Tolhurst (1947-2011)

I Introduction: texts and intentions

A theorist is some kind of intentionalist about texts if she holds one or more of the following views: First, no sequence of marks is a text unless it has the right sort of connection with an intention.1 Second, the identity of a text is determined by the right kind of connection to an author’s intention.2 Third, what a text means is what its author intended. The intentions primarily in questions are linguistic intentions, that is, intentions to produce sentences that will be understood as having certain truth-conditions. Further intentions, such as the force of given sentences, whether they are sarcastic, metaphorical, etc., also derive from the author’s intention, but presuppose the linguistic intention. This third kind of intentionalism is the topic of the present essay. The core idea of the intentionalist account is that texts are at bottom speech-act-like.3

These kinds of intentionalism are almost independent.4 The first section of this essay will argue that the second kind of intentionalism, intentionalism about textual identity, makes the third kind of intentionalism, intentionalism about textual meaning, possible and plausible. The second section will raise some difficulties for the third kind of intentionalism that arise from the differences between texts and acts. The third section will propose a somewhat weakened version of intentionalism about the meaning of a text.

II Intentionalism About Textual Identity Enables Intentionalism About Meaning

a) Utterance meaning

The linguistic meaning of an utterance is very plausibly given by the linguistic intentions of the speaker, within limits. Certainly the reference of demonstratives and the selections among syntactic ambiguities depends on the speaker. Utterances are speech-acts, individual actions by particular people, to be understood using the “intentional framework.”5 Utterances are individual events, non-repeatables.

Davidson and other intentionalists about meaning treat written texts as essentially on a par with utterances.6 But texts have features that are not found in acts, whether speech-acts or writing-acts.7 Texts are public, repeatable items. Other people can use exactly what you wrote or said to say things of their own. Because they are repeatable as the very same text, texts can be incorporated into other texts. Acts, as particular events, cannot be incorporated into other acts, cannot be repeated, and cannot be misused to say things at odds with an original.

If texts are like speech acts, though, they have an attachment to their original authors in spite of being public objects in the world, available and usable by anyone. This section defends the claim that texts are like speech-acts.

b) Skepticism about Appropriating Strings of Characters

When we think of the authority of the author’s intention over the meaning of a text as attaching to texts construed as strings of words, and therefore appropriating sequences of sentences that are the linguistic common property of everyone to mean what the first user of the string meant,8 the idea that an individual has authority over what that string means is very implausible.9 Language is common property. All the sequences of English words already exist as abstract objects, since our grammars generate them. The first person to say “My daughter goes to a pretty little girls’ school” which is many ways syntactically ambiguous, did not resolve those ambiguities in a single way for that sequence for all time.  How could someone “own” a sequence of English words? If texts are word-sequences, a reasonable view would deny the ownership or authority intuition and leave readers and interpreters to make what they choose of the text, with no constraint from the original creator of the text. That Smith said it first does not make it her property.

c) Textual identity and copying:

This essay offers a defense of intentionalism about the meaning of a text via intentionalism about the identity of a text. The key idea is a distinction between a copy of a text and a duplicate of a text.

“Copy” and “duplicate” are here used as technical terms to mark a distinction. An item A is a duplicate of item O just in case both A and O are exactly alike. So, two electrons are duplicates. An inscription A is a duplicate of another inscription O if A has the same characters in the same order as O, i.e. if O and A are instances of the same string.

Copies are one kind of duplicate. A is a copy of O if A is a duplicate of O that came into existence because of the right kind of connection to O.10 “The right kind of connection” in the case of texts, is typically that A is a copy made by an agent in order to duplicate O. Roughly and typically, copies of texts are purposely made duplicates of the original product of a text-producing act.11 So, the two electrons may be duplicates, but neither is a copy of the other.

In general, duplicates of complicated objects will be copies of such objects, given combinatorial probabilities. A duplicate of a wombat is almost certainly based on copying wombat genes, since the likelihood of a wombat coming into existence accidentally is vanishingly small.12 In the same way, a duplicate of the character-sequence that constitutes the Critique of Pure Reason almost certainly came about by some chain of copying of Kant’s original. A copy of a text is a copy of the product of an action as a product of that action.

For complicated objects, the distinction between copies and duplicates thus does not practically matter. Almost all duplicates are copies. The distinction matters in theory, though, because there is nothing in the concept of a duplicate to connect an author of one of two duplicates to the other duplicate. The word-sequences are everyone’s word-sequences. If one thinks of texts as duplicates, any claim of a real connection tie between the author of an original and a duplicate will be mystical, just as it would be mysticism to postulate a First Mother of the universe’s electrons. Duplicates need have no genuine connection with an original producer.

Notice that not all copies are intentional or have anything to do with intentions. Species of plants and animals evolve by mistakes in copying DNA sequences. Linguistic copies will generally have intentions behind them, but not always. A cat can accidentally step on the mouse and click on the “print” icon.

The thesis Tolhurst and I advanced in “On Textual Individuation” amounts to the claim that texts are identical if and only if one is a copy of the other. Suppose this is mistaken. If thesis 2) is mistaken there are two possibilities:  Either their criterion for “is the same text as” does not coincide with the usage of “text” in English, or “is a copy of” is too vague to be a usable theoretical notion. If “copy” and “duplicate” are in fact near synonyms, and if there are uses of “text” according to which duplicates of a text are instances of the same text,13 then the intentionalist can concede the point, and say that she is interested in the subset of the pairs in the “is the same text as” relation where one pair is a copy of the other. Call this subset the relation “schmextual identity.” Then the discussion below completing the move from intentionalism about identity to intentionalism about meanings will go through for schmexts. After all, the intentionalist was not claiming that the meaning of any inscription was what the first producer had in mind—that would appeal to a mystical connection.

It is difficult to see how someone could deny that some inscriptions are copies of others, in the sense described above. The most one could say is that there seems to be no easily-defined precise criterion for the “is a copy of” relation in other terms. The demand for a sharp criterion, however, would not only eliminate “is a copy of” but also “is tall,” “is a chair,” and almost every other term dealing with organisms and other medium-sized objects.14

So, intentionalism about text-identity, thesis 2) is correct, if we allow that texts may in fact be schmexts.

d) The short step to intentionalism about meaning

Intentionalism of the second kind is the thesis that texts are utterance-like, and that textual interpretation is like interpretation of a person’s utterance. This can be understood as two distinct theses: 1) Nothing is a text without the right kind15 of connection to an intention.16 2) If individual A is the same text as individual B, then either one is a copy of the other or they are both copies of a descendent of a chain of copies leading back to a third individual C. Whether an item is a copy of a text, and so an instance of that text, depends not just on what sequence of marks it instantiates, but whether it was produced so that it depends on the original “in the right way.” Being a copy requires some kind of a causal-intentional connection with that original text.17 A text can be duplicated without being copied. Two utterances of “take my wife, please” may be duplicates of one another without one being a copy of the other, and may well have different meanings.18

Acceptance of the second kind of intentionalism enables the acceptance of the third kind of intentionalism, that the meaning of a text is determined by the intention of the author. Here is the argument:

Thesis 2) is an extension of thesis 1). Further consequences of thesis 2) that do not follow from thesis 1) are: a) A text is the text it is not relative to a language, but absolutely. Your copy of The Critique of Pure Reason is really a copy of it, not a copy relative to a choice of language. b) There could be two texts that had all the same characters, yet were different texts because they were copies of or related in the “same text”-preserving way to distinct originals.

Thesis 2) affirms a kind of “trace” of the original author’s act that connects texts which are copies of the original to that original. Thesis 2) claims that since a copy is a copy of a particular act, the original intention of the act, which is a feature of that act, attaches to copies, even when not only the original intention and the act itself, but also the author has ceased to exist.

Thesis 2) is thus the key to understanding how the third degree of intentionalism, intentionalism about textual meaning, could be true even though linguistic products can have duplicates and those duplicates need have no connection to any particular author. The idea is that copies are duplicates generally made with the intention of conveying the original, which is in this case an intentional speech- or writing- action.

The intentionalist conception of the identity of a text allows the intentionalist conception of the meaning of a text. Because copying is a relation to a particular act, copies of our speaking and writing products have appropriate connection to our original intention. To review:

Step 1): A speaker or writer produces a product, an object in the world common to him and his audience.

Step 2): The linguistic intention with which the author produces the product is part of what his act is. His production is the result of his intentions—what kind of action it is, what the various purposes are, what it’s supposed to be rest on the author’s intention. The particular production-act is informed by the author’s intention. The particular product, the token, is, among other things, a product of this act.

Step 3): Copying is a relation of one product to another product as the product of an individual act. The “is a copy of” relation is transitive, but not symmetric; whereas the “is a duplicate” relation is both transitive and symmetric. Kant’s manuscript is not a copy of my first edition Critique of Pure Reason, even though it is a duplicate. Copying, the intentionalist claims, is (generally) intentionally making a duplicate of a particular product, an individual event that came about by an intention. On the other hand, any two electrons are duplicates of one another.

Step 4): An object is an instance of an author’s text only if it is a copy of that author’s text. Authors are connected to duplicates of their texts only if those duplicates are copies of the words the author spoke or wrote. Copies of copies of copies of copies are connected to the author by the transitivity of the “copy of” relation, even if the last copier does not know who he is copying. Only Zap collectors may know that Crumb is the author of “Keep on Truckin.”

Step 5):  Conclusion

Since text-ownership is not word-sequence ownership, skepticism about text-ownership is misplaced. The meaning of a duplicate of an author’s text need have nothing to do with the author’s intention, whereas the meaning of a copy of an author’s text, since it is a copy of a particular act’s product, is connected to the author’s intention in the same way19 and to the extent that the speaker’s intention determines the meaning of a speech act.20

III A Problem with intentionalism

Supposing that the Tolhurst-Wheeler view of textual identity is accepted, the opposing views on the topic of what a text means are a variety of views that make the meaning of a text depend in one way or another on the audience. Given that audiences change over time and place, and even within cultures, these audience-dependent accounts of meaning make meaning relative. The challenge to intentionalism discussed in this essay comes from the relativist side.

The major difficulty with the intentionalist account of textual meaning is that some texts incorporate other texts as components. Given this incorporation, the connection between the intention of a writer and the meaning of a text breaks down.21 This essay argues that intentionalists should adopt a kind of relativism about the meaning of such texts, but not a relativism about audiences or interpretive communities. We begin by examining three texts from the Pentateuch, the core of the Bible,22 the paradigm hermeneutical text.

a) Three Biblical texts

Consider first, Deuteronomy 27:15-26. This is a list of twelve curses. The first eleven curses have the form “A curse on him who [commits some sin]…and all the people shall say: Amen.” The twelfth curse is “A curse on him who does not maintain the words of this law [ha-torah hazzeh] by observing them. And all the people shall say: Amen.”  This is a liturgical passage being incorporated into the text of Deuteronomy. As part of this piece of liturgy, the twelfth curse is a meta-curse, and “ha-torah hazzeh” refers to the previous eleven curses. However, the sentence after the twelfth curse, Deuteronomy 28:1 is, “But if you obey the voice of Yahweh your God keeping and observing all those commandments of his that I enjoin on you today, Yahweh your God will set you high above all the nations of the earth….” “The commandments of his that I enjoin on you today” are all the commandments in Moses’ speech running from Deuteronomy 5-33. So the “but” in Deuteronomy 28:1 treats everything Moses has said in this speech as the reference of “hatorah hazzeh” in Deuteronomy 27:26.23

The reference of the complex demonstrative “hatorah hazzeh” in the original liturgical text has been understood differently when the phrase is incorporated into the larger text. Since Deuteronomy is originally a separate text,24 though, the reference in Deuteronomy is not to the Mosaic Laws found only in Exodus, Leviticus, or Numbers, but just to the laws in Deuteronomy 5-33 being enjoined. This still wider reference for “ha-torah hazzeh,” though, is understood in a later text from another scripture. In Galatians 3:10, Paul paraphrases the same passage, in Greek, as “Accursed is everyone who doesn’t follow the things written in the scroll of the law.”25 “The scroll of the law” here refers to the whole Torah, that is, the Pentateuch, and not just to the commands Moses has transmitted in Deuteronomy, since the Torah scroll is the entire Pentateuch.

Second, consider a passage from Genesis 3: In Genesis 3:8-9, Adam and Eve hear God walking in the Garden of Eden in the cool of the day, and hide. God calls to the man “Where are you?” God discovers that Adam and Eve have violated God’s prohibition against eating fruit from a particular tree, distributes punishments, and evicts Adam and Eve from the Garden of Eden. In 3:22, God observes that “the man has become like one of us, knowing good and evil. He must not be allowed to stretch his hand out next and pick from the tree of life also, and eat some and live forever.”  Prima facie, the god being quoted is a physical being with feet, is a member of a population of similar beings, and has to find out where his creatures are. In addition, people can become gods themselves by eating from a magic tree.

But Genesis 3 comes after Genesis 1, where God creates the world and other natural phenomena by ordering them into existence. Considering Genesis as a single text with a single author, God’s actions and words in Genesis 3 have to be read as figures of speech and anthropomorphisms. God is present in the garden “walking,” so to speak. When God asks questions of humans, they are rhetorical questions. God knows what Adam and Eve have done and where they are.

According to modern scholarship, the text of Genesis is a compilation of several texts by different authors with (at least) two different characters “God” with quite distinct natures.26 In the text, God is called “Elohim” in Genesis 1 and “Yahweh” in Genesis 3. Elohim is (here) a creator by fiat; Yahweh is more or less like a human with superpowers, forming people out of dirt and breathing life into them. Many passages in the books of the Bible through 2 Kings presuppose a “God of Old,” to use James Kugel’s term.27 The God of Old is not a mind reader, has more or less human size and form, can (sometimes) be safely looked upon, and is assumed to be one among several gods.28

Genesis is the product of successive compilation and editing over many centuries of multiple texts of various kinds. By the time this process is being completed, the conception of God has changed so that early texts such as Genesis 3, which presuppose the God of Old, have to be understood differently. The compilers put Genesis 1 before Genesis 3 and understand Genesis 3 metaphorically, because they take “God” in Genesis 3 to be their God, and their God is transcendent, is all-powerful, knows peoples’ inner thoughts, is not one of many gods,29 and so forth.

The third Biblical example examines some of the consequences of conceiving of the Pentateuch as the Word of God, which has come to be the understanding, certainly by the time of Ezra, after the Exile, and, given Josiah’s reception of the Book of the Law in 2 Kings 22, arguably much earlier. As God’s word, the Pentateuch must make coherent sense. Genesis 6 and 7 describe the preparations for the great flood. In Genesis 6:19-20, a verse in which God is called “Elohim,” as in Genesis 1, Noah is instructed to take two of each kind of animal. In Genesis 7:2-3 God, now called “Yahweh,” orders Noah to bring seven pairs of each kind of clean animal and one pair of each kind of unclean animal. So is Noah supposed to bring fourteen sheep or two?

According to modern scholars, these passages are parts of two distinct narratives, understood as one story by later editors. What this text meant to the last compilers, though, is something else. By Ezra’s time, the entire Pentateuch is taken to be the word of God and is becoming part of liturgy. So the apparent inconsistencies in the composite text, of which we have only given a very few,30 have come to be understood in ways that eliminate contradictions. The eleventh century commentator Rashi’s explanation, transmitting the traditional consistency-yielding understanding, is that Genesis 6:19 means “at least two,” so that Genesis 7:2-3 is a more specific instruction, but not a contradictory instruction.31

b) The meaning of the text

The three examples of incorporation of texts into other texts illustrate three ways the intention of the original author of an incorporated text can differ from the intention of the author of the incorporating text. The reference of a demonstrative can change from incorporated text to incorporating text. A sentence intended literally in the incorporated text can be meant figuratively in the incorporating text. Two texts whose authors make inconsistent claims can both be incorporated into a story making one claim.

The Bible is the text where the question of tracking authorial intention to determine meaning is especially complex. But the Bible is only the most famous example of a text which appropriates other texts. Committee documents, student papers that are pastiches of articles on Hume from the internet, and many other texts challenge the idea that an authorial intention determines meaning. Consider a few pages of an old paper on whose topic I’ve changed my mind, which get inserted into my current work with the prefatory “It might be argued that…,” and with a “…but consider the following…” at the end.

Note that these examples are not examples of texts without intentions as their cause. These examples do not speak against either of the first two aspects of intentionalism mentioned above, that nothing is meaningful unless there is an intention to communicate causally responsible for it in the right way, and that what text a text is an inscription of likewise depends on such intentions. Rather the example of the Pentateuch and its history and origins challenges the idea that the meaning of a text is determined by the author’s intention.

III Relativizing intentionalism

a) Composites as single texts

There is no question that the parts of Genesis are meaningful. Rather, the question is what is the meaning of the text as a single object of interpretation. There is a difference between a mélange of texts and a single text. To view Genesis as a single meaningful whole is to posit a single intention responsible for the text as a whole. To understand a text as a single work is to ascribe an intention to the text as a whole—to posit an author, in Foucault’s terms.32 For a text to actually be a single work with a meaning as a whole is for there to be an author whose intention is responsible for the text.33

What do we say about compilers34 as authors? If the compiler is under the impression that the components being stitched together are themselves the products of a single author, or if the compiler regards himself as constructing a narrative from disparate but consistent sources, at least the compiler thinks he is writing a single text. But what if the compiler is wrong about the intentions of his source-authors?

It seems hard to deny that Genesis is a meaningful text, even though its components are a mix of different authors’ intentions. I argue that the right position is a moderate relativism about texts and intentions. A given passage in itself means what its author intended. As a part of another author’s text it means what that other, later author intended. The user’s understanding of the incorporated text becomes the meaning of the incorporated text as an element of the user’s text.

This would give us the following picture of the career of the Pentateuch: Suppose we posit a final compiler of the Pentateuch, call him Ezra. If, as in fact happened with the Pentateuch, Ezra’s text is itself paraphrased, cited, and in other ways re-used with basic linguistic intentions substantially the same as Ezra’s, and this continues for twenty-three hundred years, it would make sense to say that, for the people understanding the text as Ezra did, the text means (roughly) what Ezra intended.35

On the other hand, there is some inclination to say that Genesis as a whole does not mean anything at all, because there is no linguistic intention informing the text as a whole, given the compilers’ misapprehension about their sources. From the point of view of modern scholarship, Ezra and other redactors misunderstood the texts they were incorporating into their compilations and redactions. So, an argument could be made that Genesis as a whole does not mean anything, because it is not a whole.

This is what we might say about Philosophical Instigations, a collection of slips of paper in the nachlass of an important philosopher, Wittstein, that were taken to be paragraphs he wrote as expressions of a new philosophical theory, but which in fact were his collection of student in-class responses to the repeated assignment, “Write something short and interesting about language.” Even though admirers of this philosopher have constructed interpretations of the collection which make it provocative and interesting, so that now there are specialists in the subject, one is tempted to say that the text as a whole has no meaning, because it is not really a text at all. One would continue to say this even though a series of editors had corrected what they had come to regard as typographical errors or misstatements, so that the “text” as it existed hundreds of years after the death of the philosopher was rather different from the original pile of student paragraphs.

d) Compilers’ and transmitters

One problem with saying that such texts are meaningless is that other texts whose intention is to transmit the meaning of these texts are attempting to transmit the meaning of a nontext, an aggregate with no actual meaning, because no actual informing intention. In the case of the Pentateuch, such secondary texts abound.36 Consider for instance the Targums, paraphrases of the Hebrew into Aramaic. These texts are paraphrases, rather than strictly speaking translations, in that they do not, even in Targum Onkelos, the most “literal,” try to convey the Hebrew word-for-word. What they present is the text as it has come to be understood.

If an actual authorial intention is required in order for the text to have a meaning as a whole, then, since the texts the paraphrase paraphrase are meaningless, so would be the paraphrases, there being (actually) no unified authorial intention. If the posited author was just a mistake, are the paraphrases meaningless also?  The same consequences would follow about the Wikipedia and the Stanford Encyclopedia of Philosophy articles on Wittstein.

Here is what an intentionalist should say: There is indeed no actual authorial intention behind the Pentateuch, just as there is no philosophy of the later Wittstein. However, the compiler of the Pentateuch and the discoverer of the pile of papers either posit such an authorial intention or purport to do so. They are mistaken, but this does not mean that their texts are in fact not single texts at all.37 If they think or intend to lead the reader to think that there is an authorial intention in the pieces they are compiling, then their texts are single texts. Then of course the translators and paraphrasers are intending to convey the meaning of the compiled or discovered text, and so these are also in fact single texts.

The actions of compilers, translators, and paraphrasers have unified authorial intention, albeit one directed at non-existent entities, the unified author of the sources. Actions can obviously be directed at non-existent entities. The king of Moab can sacrifice his first-born in order to placate Chemosh (2 Kings 3:27). I once left cookies and milk for Santa Claus. Texts can have authorial intentions to transmit authorial intentions that do not exist.

e) Conclusion:

Intentionalists who claim that the meaning of a text is to be derived from the author’s intention must acknowledge that a kind of relativism is in fact correct. When texts are incorporated into other texts, and the authorial intention about the incorporated text differs from that of the incorporating text, the meaning of the incorporated text in itself is different from the meaning of the text as part of the larger text.

This does not amount to the view that the meaning of a text is relative to a reader or to an interpretive community. Christians, according to modern scholarship, are just mistaken about what Isaiah means in Isaiah 7:14-16. The text is a reassurance to Ahaz that in a relatively short time, namely by the time a baby has grown up enough to be morally responsible, his enemies will have disappeared.

However, when the author of Matthew 1:22-23 writes “Now all this took place to fulfill the words spoken by the Lord through the prophet: The virgin will conceive and give birth to a son and they will call him Immanuel,” what he means is given by his authorial intention to point out a prophecy about Jesus. When later texts allude to Matthew, using “Immanuel” to refer to Jesus, and when “Immanuel” comes to be treated as another name for Jesus, that is who their name “Immanuel” refers to, and who Kant’s parents had in mind when they named him.