I. The Intentionalist Debate and Test Cases
The debate over authorial intention and artistic interpretation (often called the intentionalist debate) has generated a voluminous literature in philosophy and literary studies. Various positions have been proposed and the debate is ongoing. In this paper, I focus on Anglo-American analytic philosophers’ take on this issue. Among the most discussed positions in the analytic philosophy of art are actual intentionalism, hypothetical intentionalism, anti-intentionalism, and the value-maximizing theory.1 My aim here is to expose serious problems with these theories by testing them against a particular literary work, ‘The Ornamented Zither’, composed by the ancient Chinese poet Li Shang-yin.2
Using particular works of art as test cases is common practice in the philosophical literature. It has proven to be an effective way to provoke discussion and strengthen/reject philosophical points. As a useful guide, a list of test cases used in analytic theories of interpretation is provided in Hans Maes’s discussion of the interpretation of visual art.3
What I attempt to do here is to assess the mainstream positions on interpretation by testing all of them against the same test case, which I take to be a typical example of interpretation. The result will be that all these theories have potential inadequacies. If my conclusion holds, supporters of the theories criticized will need to surmount the difficulties I raise if they want their accounts sustained in the intentionalist debate.
‘The Ornamented Zither’ is a paradigmatic case of a text demanding interpretation because it has attracted great critical attention on account of its ambiguity. Interpretation is typically called for when ambiguity arises, and theories of interpretation remove ambiguity by telling us which literary meaning or meanings to adopt. Choosing a poem as the subject of discussion is also a good way for us to reconsider the intentionalist debate, since the debate officially started in the discussion of poetry when William K. Wimsatt and Monroe C. Beardsley published their seminal paper ‘The Intentional Fallacy’.4
To facilitate discussion, I briefly introduce Li Shang-yin and ‘The Ornamented Zither’ before testing mainstream theories of interpretation against the poem. I will summarize the results of the test in the final section of this paper.
II. Li Shang-Yin and ‘The Ornamented Zither’
Li Shang-yin (813?–.858) lived at a time when the once-powerful Tang dynasty (618–907) had begun to decline as a result of the usurpation of power by the eunuchs, the feud between political factions, and the rise of provincial warlords.5 Throughout his life, Li allied with different patrons for official positions, but his career was full of frustrations, partly because of the political turbulence in his time. Though far from a successful official, Li left behind a rich body of poetic compositions, which has made him one of the most acclaimed poets in the history of Chinese literature.
Although Li Shang-yin wrote quite a lot about political and historical events, his best-known works are often referred to as his ‘ambiguous poems’. Written in a highly ornate style and seemingly about amorous affairs, these poems are prominent for their obscurity and ambiguity, in part because many of them are untitled, and in part because the abundant uses of allusions and metaphors render the already-fussy contents more perplexing and difficult to comprehend.
‘The Ornamented Zither’ is Li Shang-yin’s most famous ambiguous poem. Even though it has a title, this sheds little light on the interpretation of the work itself, since the title is simply taken from the first two characters of the poem’s first line. Below is a literal translation of the poem by James J. Y. Liu:
‘The Ornamented Zither’
The ornamented zither, for no reason, has fifty strings;
Each string, each bridge, reminds one of a youthful year.
Master Chuang was confused by his morning dream of the butterfly;
Emperor Wang’s heart in spring is entrusted to the cuckoo.
In the vast moonlit sea, pearls shed tears;
At sun-warmed Blue-field, jade engenders smoke.
This feeling might have become a thing to be remembered;
Only, at the time one was already bewildered and lost.6
There are at least four allusions in this poem, appearing between line three and line six.7 Line three alludes to a famous allegory in Chinese philosophy: Zhuangzi’s dream of the butterfly, in which the Daoist philosopher Zhuangzi dreamt of being a butterfly.8 On awakening, the philosopher began to wonder whether it was he who dreamt of the butterfly or the butterfly who dreamt of him. Line four alludes to the myth in which Emperor Wang, who had a love affair with his prime minister’s wife, was overcome with shame and transformed into a cuckoo after he died. In the Chinese language, ‘heart in spring’ means the desire for sex. The allusion in line five is extremely obscure. It might allude to an old Chinese legend that oysters would contain pearls at full moon that disappear when the moon wanes, to the story in which a mermaid shed tears of pearls on a jade plate after being hosted by a human, or to the expression ‘pearl left in the vast sea’, which is a metaphor for someone whose talent is not appreciated. The allusion in line six is also obscure and might be multilayered: it might allude to another poet’s remark that compares the beauty of poetry to Blue-field (a famous place in China that produces jade) to mean that good things are always beyond one’s reach. Two other possible sources of allusion are the story in which a celestial being rewards a kind man by giving him pebbles that grow into jade, and the story in which the daughter of an ancient king, going by the name of Jade, died of grief and vanished into smoke after her spirit appeared to her father, who disapproved of her marriage.
Numerous interpretations have been proposed by literary critics to explain this enigmatic poem. According to a canonical analysis of the poem offered by Liu, commentators’ interpretations can be sorted under five headings in terms of their nature. I am adding Liu’s own interpretation as the sixth on the list. Each of these interpretations has variations owing to differences in details, but I will set aside the ontological question of whether these variations count as different interpretations or only as versions of interpretations. So as not to complicate matters, I will discuss only the variations of one of the canonical interpretations of the poem in question. Below I summarize the six interpretations to be examined.9
Interpretation I: the poem is about amorous love. Notably, this interpretation has three variations. (A) The poem was written for a maid called Chin-se (the transliteration of ‘ornamented zither’ in Chinese), believed by some commentators to be a family member of Li Shang-yin’s first patron. The four middle lines of the poem describe the four kinds of music produced by the zither and played by Chin-se (more on the four kinds of music later). (B) The poem was written to record a frustrated love affair with some woman encountered at some point in the poet’s life. The four middle lines reflect the poet’s bewildered, lost mind, as these lines themselves are vague and confusing. (C) The poem was written for the death of two sisters who had a love affair with Li Shang-yin. This interpretation is based on historical records and textual clues found in other poems written by Li. References to phoenixes and luan-birds (another kind of Chinese mythological creature) are frequent in Li’s poems, and some commentators believe that these have something to do with two particular court entertainers whose names made reference to said bird species. It is further speculated that the two sisters committed suicide before several court entertainers were executed in 839 for political intrigues. According to this interpretation, Emperor Wang’s story in line four refers indirectly to the suicide of the two girls since both events are about death. The mentions of pearls and jade in line five and six, respectively, should be considered together in light of the mermaid story: the poet gave the two sisters a jade plate to shed tears on as a gift of parting before they took their lives. This last part of the interpretation is supported by the fact that in the poem titled ‘Peonies Damaged by Rain at Hui-chung’ Li writes that ‘[t]ears sprinkled on jade plates repeatedly hurt one’s heart’.10
Interpretation II: the poem simply describes the four kinds of music played on the zither. This interpretation is based on extant records about ancient Chinese music, in which it is reported that the zither can produce music that is leisurely, plaintive, pure, and harmonious, corresponding to the moods suggested by the four middle lines of the poem as follows: the dream of the butterfly (leisurely), Emperor Wang’s regret (plaintive), pearls and tears (pure), and the fine scenery in Blue-field (harmonious).
Interpretation III: the poem was written to show sorrow for the death of the poet’s wife. In ‘Chamber Music’, another poem apparently written to mourn his wife, Li Shang-yin writes that ‘[t]he ornamented zither has lasted longer than the person’.11 This interpretation finds strong support in the first couplet of the poem under discussion. According to legend, the zither is said to have originally had 50 strings but it was broken into 25 strings on the order of a god because the music played by the zither was unbearably sad.12 In Chinese, ‘broken strings’ means the death of one’s wife. The first couplet thus tells us that the poet thought of his deceased wife upon seeing the zither. Line three alludes indirectly to the story in which Zhuangzi ‘mourned’ his wife by pounding a tub and singing. Line four suggests that the poet was living where Emperor Wang once ruled. The lines mentioning pearls and jade are said to describe the eyes and complexion of the poet’s wife.
Interpretation IV: the poem laments the poet’s unsuccessful official life. It should be noted that the Tang dynasty declined partly because of the Niu–Li factional strife that lasted for almost 40 years. Li Shang-yin’s official life was affected by this strife to a substantial extent. Ling-hu Chu – Li’s first patron – and his son Ling-hu Tao belonged to the Niu faction. After the death of Ling-hu Chu, Li Shang-yin married the daughter of a military governor affiliated with the Li faction. This caused the younger Ling-hu’s resentment against Li Shang-yin and his subsequent refusal to get Li official posts when he himself became politically influential. Based on these historical facts, the present interpretation goes as follows: the allusion to Zhuangzi’s dream refers to constant changes in the world; the allusion to Emperor Wang suggests Li Shang-yin’s powerlessness over the factional strife; the line about pearls alludes to Pearl Cliff, where the leader of the Li faction was banished; and the line about jade means that Ling-hu Tao’s successful political career is beyond ordinary people’s reach.
Interpretation V: the poem is an introduction to the poet’s collection of his own poetry. It is known that Li Shang-yin edited two collections of his prose works. It should not be surprising that he would want to do the same for his poetry. According to this interpretation, this poem was to be placed at the beginning of the collection as an introduction to the poet’s works ‘and thus more generally as a poem about the poetic art’.13 The first couplet alludes to the legend of how the zither’s strings were broken in half and thus suggests the expressive power of art. And the subsequent couplets point to the problem of indeterminacy of poetic meaning: ‘[W]hile the compression of his poetic language leads us to infer a latent intensity of emotion, that same compression obliterates the particularity of reference, and in the end the exact source and nature of this feeling eludes any attempt – by poet or reader – to pin it down once and for all.’14
Interpretation VI: the poem shows the poet looking back at his life. This interpretation claims that the first couplet tells us how the poet begins his retrospection: just as a zither has a given number of strings, so does a man have a given number of years to live. The second couplet suggests that life is like a dream and that one can hardly know which is real: Zhuangzi or the butterfly? Emperor Wang or the cuckoo? The third couplet laments past loves: tears were shed when those you loved must leave you, and they vanished like the smoke engendered by jade. The final couplet, as Liu concludes, expresses the feeling that ‘[a]ll these experiences of the past might have become memories to be cherished, only, even at the time they occurred, one was already bewildered and could not be sure they were real’.15
With these popular interpretations in mind, let us test analytic theories of interpretation against ‘The Ornamented Zither’.
III. ‘The Ornamented Zither’ and Analytic Theories of Interpretation
III.1. Anti-intentionalism
Anti-intentionalism, represented by Beardsley, is the doctrine that in most cases linguistic conventions determine a single right interpretation of a work.16 These conventions are semantic, such as word meanings recorded in dictionaries, and syntactic rules. According to the anti-intentionalist, they generate internal evidence for the meaning of a work. External or extratextual factors, such as evidence of the author’s intention and the history of the creative process, are considered irrelevant to the shaping of textual meaning. To assess this position, there are a couple of nuances to consider.
First, what is internal to a text is further defined as ‘all the literature which is the source of dictionaries’ and in general ‘all that makes a language and culture’.17 This deflects the objection that anti-intentionalism falls short of handling allusions and intertextual references.
Second, the anti-intentionalist allows use of intermediate evidence, which is ‘evidence about the character of the author or about private or semi-private meanings attached to words or topics by an author’, and this is because ‘the meaning of words is the history of words’.18
Third, it is not true that anti-intentionalism can only deal with the literal meaning of a linguistic unit. Beardsley has demonstrated how an anti-intentionalist can infer from a linguistic unit’s literal meaning to its implicit meaning.19 For example, the word ‘sea’ designates a large body of water but it also ‘connotes certain other characteristics, such as being sometimes dangerous, being changeable in mood but endless in motion, being a thoroughfare, being a barrier, and so on’.20 The connotations of a word are constrained by the word’s explicit, literal meaning, and textual contexts can tell us which is the correct connotation of the word under consideration. The same manoeuvre works for what a sentence implies or suggests.
It would seem that, among the popular readings of ‘The Ornamented Zither’, the anti-intentionalist approach outlined above makes available interpretation IB (love affair with a woman), III (mourning the poet’s wife), VI (looking back at one’s life), and, controversially, V (introduction to collected poems) if knowledge of Li Shang-yin’s oeuvre is taken as internal evidence, as the expanded notion of such evidence mentioned earlier might have suggested. All other interpretations are based on facts external to the poem.
Now, the anti-intentionalist faces the problem of underdetermination. It is not clear which interpretation among the four wins out. They are all supported by internal evidence of the poem, and there does not seem to be a non-arbitrary way to adjudicate on them. Apparently, the resources provided by anti-intentionalism are not enough to break such a deadlock if the aim is to find the correct interpretation.
Underdetermination can happen even within interpretations that share the same guiding thought. Take interpretation IB, for example. Ji Xiaolan, a literary scholar and proponent of this interpretation, maintains that ‘[a]t first there was someone he loved, then there were obstacles, so he wrote this poem to recall her. The four middle lines are vague and confusing – this is what is meant by “bewildered and lost”. It has no deeper meaning.’21 Liu criticizes interpretation IB as ‘too easy a way out’.22 Suppose that one modifies the present interpretation by specifying the emotions implicated by the four middle lines. For instance, it could be maintained that the line about Zhuangzi is suggestive of the dreamy quality of amorous love that emerges from infatuation, that the lines about Emperor Wang and pearls signal regret and sadness, and that the line about Blue-field conveys joy and leisure. Call this interpretation IB*. The question now is whether or not internal evidence alone can adjudicate between IB and IB*.
I think that the answer is negative. A confused mind can be embodied by vague and confusing allusions. On interpretation IB, the poet is confused and bewildered because he is not able to elucidate the sentiments entangled in his mind. On interpretation IB*, however, the poet is confused and bewildered because he had experienced different, even incompatible sentiments. Both interpretations are supported by textual evidence, and textual evidence alone is not sufficient in discriminating between the two.
Supporters of anti-intentionalism might object that their idea of what constitutes internal evidence is richer than my reading implies and that I therefore misrepresent their position at a fundamental level. If internal evidence is indeed public and all that makes a language and culture, then the scope of this sort of evidence is certainly wider than the linguistic conventions relevant to the words that form the text being interpreted. In that case, however, all the interpretations discussed in Section II beyond the four suggested earlier will become available to the anti-intentionalist, since they can be reached from different pieces of evidence such as gossip about Li Shang-yin (IA), historical records about political intrigues in 839 (IC), extant records about ancient Chinese music (II), or the biography of Li Shang-yin (IV).
It seems clear that adopting this richer notion of internal evidence conflates internal and external evidence, which makes futile the distinction between the two. Michael Wreen’s recent article on Beardsley reveals that the internal/external distinction is problematic as it is couched in its first appearance. After all, evidence of the author’s intention or her creative history is also publicly accessible. According to Wreen, the internal/external distinction is better treated as one about what information is relevant/irrelevant to work interpretation. Relevant information is internal in the sense that it can be gleaned by a direct inspection of the work; irrelevant information is external in the sense that it cannot be gleaned by doing so.23 All in all, the discussion so far at best clarifies confusing terms in the founding document of the anti-intentionalist literature, which is certainly beneficial to Beardsleyan scholars. But it does not resolve the problem of underdetermination in the case of ‘The Ornamented Zither’.
The anti-intentionalist could claim that ‘The Ornamented Zither’ is ambiguous in nature, as anti-intentionalism accepts that there can be such cases on condition that they are few and far between. As mentioned at the beginning of this paper, however, ‘The Ornamented Zither’ is a typical case of interpretation. The move suggested here would make every typical case of interpretation unresolvable, which runs counter to the stated claim that ambiguous texts are rare.
There is a further problem with anti-intentionalism: in many cases reference to authorial intention seems inevitable. Here is David Davies:
While one might initially think that Wimsatt and Beardsley subscribe to a textualist conception of the literary work, it is clear that, in allowing evidence as to the ways in which the author of a text employs language in other contexts, they are taking a literary work to be a particular text generated by a particular author, and thus as generated in a particular context. Thus their position is a kind of contextualism.24
Textualism identifies a literary work with a text. Work-meaning is thereby textual meaning – or word sequence meaning, as some call it – determined by ‘the operative syntactic and semantic (including connotative) rules of the specific, time-indexed, language in which those words are taken to occur’.25 Contextualism identifies a literary work with a text used on a particular occasion. On this view, it is plausible to see work-meaning as utterance meaning, and the context of utterance contributes partly to the shaping of meaning.26
Allusion is another case in point where the anti-intentionalist seems unable to eschew reference to the author’s intention and the contextualist sentiment. For the anti-intentionalist, whether a poem alludes to another work is determined by whether that allusion makes sense in the poem under consideration if the author is thinking about it.27 But ‘alluding to the work of another writer is something an author does in a poem, rather than something the poem itself can do purely in virtue of its own internal resources’, and this ‘requires that language be intentionally used in a certain way by an author or speaker’.28 It is thus difficult to see how anti-intentionalism can accommodate allusions and intertextual references without mobilizing the notion of authorial intention in such a case as ‘The Ornamented Zither’, where allusions and intertextual references abound.
Let us now consider another theory of interpretation that improves on anti-intentionalism. The value-maximizing theory shares the guiding principle with anti-intentionalism that linguistic conventions play a key role in interpretation. The maximizer, nevertheless, recognizes the problem of underdetermination exposed earlier. In other words, the maximizer takes a thin notion of convention, while the anti-intentionalist invokes a robust one.29 A normative criterion is then introduced to adjudicate between interpretations supported by textual evidence: we should choose the interpretation that can maximize the value of a work. Since there can be more than one way to maximize a work’s value, a work can admit of multiple interpretations that are equally acceptable.30 Indeed, it has been suggested that ‘maximizing’ be better put as ‘enhancing’ so as to avoid the wrong impression that the position in question champions critical monism, the position that there is only a single right interpretation of a work.
In so far as it derives from anti-intentionalism, the maximizing view faces the same problem of explaining intentional use of language. Aside from that, there is also the problem of measuring aesthetic value.31 The anti-intentionalist requires that the interpreter seek an interpretation that makes best sense of the work based on textual evidence. The maximizer acknowledges this but adds that there can be multiple interpretations that satisfy the present requirement and that we should exclude those that do not enhance a work’s value. To put things more accurately, conventions alone offer us interpretations that make good sense of a work but never those that make best sense of a work.
As I will argue, if an interpretation is already one that makes good sense of a work, it is naturally value-enhancing. Each of the popular interpretations of ‘The Ornamented Zither’ attempts to make best sense of the text, especially the four middle lines that appear enigmatic. And, seemingly, they are equally good interpretations even if none of them is the best. Under the constraints of conventions, it is plausible to construe a good interpretation as one that is comprehensive, internally consistent, and sufficiently supported by textual evidence. Such an interpretation enhances our appreciation of a work because it helps us understand the work better. If so, it does make sense to view this sort of interpretation as value-enhancing. Not only does the actual intentionalist accept this point;32 it also seems to be one of Beardsley’s convictions: ‘The work is an object capable (presumably) of affording aesthetic satisfaction. The problem is to know what is there to be responded to; and the literary interpreter helps us to discern what is there, so that we can enjoy it more fully.’33 Now consider a counterexample to my argument about value-enhancement in which the interpretation of a work ascribes negative aesthetic properties to it. For example, an interpretation that renders a work uninteresting but is compliant with the text is not value-enhancing. However, having aesthetic properties of negative valence is but one ground for aesthetic condemnation of a work.34
Some might argue that the maximizer could introduce a narrower conception of aesthetic value, which does not imply that all good interpretations result in the enhancement of a work’s value. Philosophers adhering to this perspective will claim that coherence, for example, is still a fairly neutral thing and that a coherent interpretation of a work is thus not value-enhancing. Nevertheless, the maximizer seems to support a broad conception of aesthetic value, including cognitive engagement with literary works.35 In sum, if we accept that the cognitive task of ascertaining a work’s meaning enhances its value by enriching our appreciation of that work and that constructing a comprehensive, coherent, and text-complying interpretation facilitates literary understanding, then it follows that such an interpretation is value-enhancing.
III.2. Intentionalism
There are at least two breeds of intentionalism by virtue of the nature of intention: actual intentionalism and hypothetical intentionalism. I start with the former.
Actual intentionalism has a formidable list of supporters. Though this position has different versions, they share the claim that the author’s intention plays a key role in settling the meaning of a work.36 This claim is normally based on both the assumption that the meaning of a work is uniquely determined by the author’s intention, and the assumption that interpreting a work is parallel to interpreting a conversational utterance.
An immediate difficulty for actual intentionalism is that the author’s intention is often not available.37 In the case of ‘The Ornamented Zither’, it is impossible to ask Li Shang-yin about his intention since he is long dead, and documentary evidence in that respect is simply not forthcoming. The case of ‘The Ornamented Zither’ is typical: it is often the case that evidence for the author’s intention beyond what is in the work is not available because it is impossible or difficult to obtain. A simple test is to look to the vast body of works by dead authors, particularly those about whom little has been written. It turns out that the interpretative policy recommended by the actual intentionalist is highly inapplicable, especially in resolving a typical case of interpretation.
It might be objected that actual intentionalism can be sustained solely by the metaphysical claim about the author’s intention being meaning-shaping. Nevertheless, the history of the intentionalist debate shows that the normative question of how an interpreter should proceed lies at the heart of the debate, or at least is a necessary part of it. A theory of interpretation cannot only offer metaphysical tenets about meaning; it must also provide a sensible answer to what the interpreter should do.
One way for the actual intentionalist to address the unavailability argument is to say that the critic should remain agnostic about the meaning of the work; insufficient data or evidence often makes it rational to refrain from firm fixation of belief. The critic can survey and imagine possible meanings but not take any leap of faith. And this is what the critic should do in the case of ‘The Ornamented Zither’.
An immediate concern with the above rejoinder is that it does not reject the conclusion that evidence for authorial intention is generally unavailable. The result is then that the actual intentionalist should remain agnostic in most cases of interpretation, which seems implausible. Note that embracing agnosticism about what a work means in this context is different from making a hypothesis about it. In the latter case, the interpreter offers an interpretation she firmly believes in, while at the same time accepting the possibility that it can be shown by some additional evidence to be problematic or even false. By contrast, the agnostic will not have any preferred interpretation and will accept the indeterminacy of meaning. Earlier I emphasized that a theory of interpretation must provide a sensible answer to what the interpreter should do. If a theory of interpretation tells the interpreter to suspend her judgement most of the time, it does not seem to be a useful strategy for interpreting works.
Additionally, empirical evidence shows that, at least in the case of Li Shang-yin’s ambiguous poems, critics or commentators trying to recover his intentions do not embrace agnosticism. In his discussion about Li’s ambiguous poems, Liu identifies three schools that are most influential in studying and deciphering these poems.38 Scholars belonging to the first school see Li’s ambiguous poems as veiled references to his patrons (or would-be patrons); the second school views them as alluding to clandestine love affairs; the third school thinks of them as satires upon political situations in Li’s time. On Liu’s view, all these three schools seek the poet’s creative intentions that caused his poems. The practical lesson from this empirical fact is that, when confronted with interpretative disputes where decisive evidence is absent, what the critic adopting actual intentionalism would do is not stay agnostic but proffer an interpretation she thinks to be what the author had in mind. I am not using this example to say that what critics are doing defines interpretative norms, since that will logically be a fallacy. Rather, my point is that, if a theory advises critics to do something most of them are not likely to do, then we should be sceptical of the advice. An important claim made by the modest version of actual intentionalism that is related to the present objection to the unavailability of authorial intention is that the text is the best evidence for the author’s intention.39 Therefore, the interpreter can rely primarily on the text to figure out the author’s intention. This claim is plainly false. If the text of ‘The Ornamented Zither’ were really the best evidence for Li Shang-yin’s creative intention, commentators on the poem would have been able to make out what he intended it to mean – namely, the unique meaning of the poem. The fact that the poem in question supports multiple interpretations shows that the text of the poem is not the best evidence for Li’s intention.
Alternatively, the actual intentionalist could say that the text of ‘The Ornamented Zither’ indeed is the best evidence of the poet’s intention, because it shows that Li Shang-yin intended the poem to be multi-interpretable. As mentioned, Li Shang-yin is noted for writing ambiguous poems, most of which are untitled.40 It is then no surprise that ‘The Ornamented Zither’ could have been intended as one such poem.
This reply will not do for the reason that it conflates ‘intending an ambiguous text to mean a specific meaning p’ with ‘intending an ambiguous text to be multi-interpretable’, hence making every ambiguous text the product of the latter sort of authorial intention. The typical case of interpretation is the case in which the meaning of a text is not salient, and the present reply implies that in every such case the author intends her text to be multi-interpretable, which is obviously false.
The modest intentionalist might object that I misrepresent the best-evidence claim in a fundamental way: what she actually means by that claim is not that the work taken completely independently could serve the type of role I mentioned but that it has to be seen against the background of some relevant general knowledge about the context of creation and the author. On this reading, the claim in question maintains that the work itself serves as a kind of clinching evidence in the whole.
I suspect that this is the correct reading of the best-evidence claim. We should probably examine more closely why and how this claim was made. As early as his defence of actual intentionalism, Noël Carroll claims that authorial intention is a purpose that manifests in the artwork and ‘regulates the way the artwork is’ (this is dubbed the neo-Wittgensteinian view). He also adds that authorial intention ‘is discoverable by the inspection and contemplation of the work itself’.41 In his subsequent defence of modest intentionalism, he rejects a frequently raised objection to actual intentionalism: it diverts the audience away from the proper object of interpretation. Carroll reiterates this objection and offers his reply:
Instead of focusing on the text, intentionalism sends the reader outside the text, searching for the author’s intention – perhaps in the archive where his private papers are stored. This criticism, however, is misguided here, since the modest actual intentionalist freely admits that what an utterer, artist, or author intends to say or mean is the utterance or artwork itself. Modest actual intentionalism is not an injunction to root for authorial meaning in hidden places. Generally, we find authorial intention expressed in the artworks in question.42
It is very clear from the citations above that the text is the best evidence for the author’s intention for the reason that very often such intention is discoverable in the artwork itself. But being discoverable does not imply that it is always easy to make out the intention hidden in the work. This observation is confirmed by Carroll’s claim that the audience can discern authorial intention in the work after being informed of what that intention is when it is not conspicuous to the appreciator.43 But this manoeuvre reverts to the problem of the unavailability of authorial intention. Note that I am not denying that the text can be evidence for authorial intention. Certainly it can be taken that way. However, this sort of evidence does not provide very useful guides in cases like ‘The Ornamented Zither’.
To continue the debate, the modest intentionalist could persist that a more nuanced account of interpretative disputes can be given and that there is no reason why she could not say that some pieces of the contextual and textual evidence cited in the case of ‘The Ornamented Zither’ are better ones of the poet’s actual intention than some of the others. There is perhaps no definite methodology for identifying such pieces of evidence, but still this option is available to the proponent of modest intentionalism.
Indeed, some modest intentionalists such as Robert Stecker lay special stress upon contextual evidence, including the author’s oeuvre, the sociopolitical conditions in the author’s time, the relevant art traditions, the author’s biography, and so on.44 All these along with textual evidence are used in the canonical interpretations of ‘The Ornamented Zither’ introduced earlier. Suppose that the modest intentionalist has devised a criterion for identifying which piece of evidence is better than others. The author’s oeuvre, say, matters most in interpretation, so ‘The Ornamented Zither’ should be interpreted as mourning the death of the poet’s wife, since the evidence in question reliably indicates the poet’s intention in the poem.
But this response begs the question. The problem is exactly how the criterion under discussion can be thought out in a non-arbitrary way. Why does Li’s oeuvre, for example, matter more than facts about his political career in indicating his intention? By extension, why do authors’ oeuvres matter more than facts about their life in disclosing their intentions? At least in the case of ‘The Ornamented Zither’, it is difficult to see such a criterion (or criteria) forthcoming in adjudicating on the contextual evidence cited for consideration.
In general, perhaps, the actual intentionalist could loosen the epistemic requirement of recovering the author’s intention in the typical case of interpretation where extratextual evidence of such intention is not available. It could be held that in such a case the best we can do is make the best hypothesis about the author’s intention based on whatever evidence we can gather. The epistemic obstacle does not obliterate the fact that it is still the author’s intention that we are after. In practice, however, this move renders actual intentionalism indistinguishable from its intentionalist rival – hypothetical intentionalism, whose leading figure is Jerrold Levinson.
On Levinson’s view, work-meaning is determined by the appropriate audience’s best hypothesis about the author’s semantic intention in terms of an evidence base that excludes knowledge of such intention.45 To clarify, an important distinction made by Levinson is that between semantic intention and categorial intention:
An author’s intention to mean something in or by a text T (a semantic intention) is one thing, whereas an author’s intention that T be classified or taken in some specific or general way (a categorial intention) is quite another. Categorial intentions […] govern not what a work is to mean but how it is to be fundamentally conceived or approached.46
Strictly speaking, the intentionalist debate is mainly about the relevance of semantic intention. If one agrees with what Levinson says here about categorial intention, then there is no inconsistency in accepting and consulting the author about this sort of intention. In other words, one can be an actual intentionalist about categorial intention for the reason just given while being a hypothetical intentionalist (or a theorist of other stripes) about semantic intention. Apparently, Levinson’s view about categorial intention derives from Kendall Walton’s influential account of categories of art, in which Walton claims that the correct categories of an artwork fix what aesthetics properties it has and that the correctness in question is principally determined by the author’s intention.47 Unless specified otherwise, the word ‘intention’ in the remainder of this paper refers to semantic intention.
Now, the reason for restricting the evidence base for the appropriate audience in the way suggested earlier is because, according to Levinson, there is an implicit contract between author and audience that the former’s intention not be consulted.48 Though this claim is controversial, let us accept it for argument’s sake. One problem with hypothetical intentionalism is that different audiences using the same evidence base stipulated as above are likely to make competing hypotheses about the author’s intention, with none of them uniquely best. In other words, the evidence base used by hypothetical intentionalism underdetermines the best hypothesis about the author’s intention.49
Let us consider ‘The Ornamented Zither’ again. According to Liu, proponents of interpretation I to V are all aiming at the best hypothesis about Li Shang-yin’s intention, because all of them attempt to find out what occasioned the writing of the poem.50 All these hypotheses are made against the same textual evidence (the text of ‘The Ornamented Zither’) and extratextual pointers (biographical facts, Li Shang-yin’s oeuvre, the political background of the late Tang dynasty, knowledge of Chinese music, and so on). But these hypotheses differ in the weight they attach to various respects of the evidence considered. Interpretation I hypothesizes that Li intended the poem to express his feelings about an affair in the past based primarily on some gossip from his life; interpretation II hypothesizes that Li intended the poem to describe the beautiful music played by the zither in light of our knowledge of ancient Chinese music; interpretation III hypothesizes that Li mourned his wife by writing the poem on the basis of what we know about his marriage and another poem written by him; interpretation IV hypothesizes that Li wrote the poem to lament his official life in view of biographical facts about him and the political climate of the late Tang dynasty; and interpretation V hypothesizes that Li intended to write about poetic art, given what we know about his oeuvre.
The root of the underdetermination in question stems from the fact that there seems to be no objective criterion for determining which respect(s) carries greatest weight in making the hypothesis epistemically best. Does the fact that ‘Chamber Music’ mentions the ornamented zither make it more likely that Li Shang-yin intended ‘The Ornamented Zither’ to mourn his wife? Or does the gossip that Li was on close terms with some maidservant in his patron’s household make it more likely that he wrote the present poem for her? The fact that commentators’ opinions are divided shows that different readers are bound to make different hypotheses that are equally best, even if they adopt the same evidence base.
At this juncture, proponents of hypothetical intentionalism may object that it is possible for them to give different weight to different textual, contextual, and authorial evidence, and to make meaningful distinctions between them. My criticism of hypothetical intentionalism assumes that its proponent must treat all textual, contextual, and authorial evidence with equal relevance. If this counter-reply makes sense, it is possible to reach a single right interpretation in my case study of ‘The Ornamented Zither’.
I am sceptical of this counter-reply, because it does not seem to be what the hypothetical intentionalist has in mind with respect to the relevant evidence in making the best hypothesis. Levinson says explicitly that literary meaning is
a function of and constrained by […] the potentialities of the text per se together with the generative matrix provided by its issuing forth from individual A, with public persona B, at time C, against cultural background D, in light of predecessors E, in the shadow of contemporary events F, in relation to the remainder of A’s artistic oeuvre G, and so on.51
The above construal of hypothetical intentionalism tells us that the position under scrutiny does not put particular emphasis on specific evidence or information such as A or B or the like. One important assumption for Levinson is that work-meaning is analogous to utterance meaning, the meaning a linguistic vehicle ends up conveying in its context of utterance. It is difficult to identify which condition(s) is (or are) more crucial in locating utterance meaning, and normally there are no rules for achieving this. What is conveyed on a particular occasion of use of language is basically the result of the total evidence or information combined together, and based on what is available to the hearer, she makes a general assessment of it, trying to apprehend what is said.
As far as the problem of underdetermination is concerned, Levinson proposes at least three solutions. The first is to disjoin all competing hypotheses into a grand reading.52 This approach has been questioned by Gregory Currie,53 who advocates such a disjunctive hypothesis on behalf of monism only if the disjunctive reading is better than every disjunct. However, this cannot be the general case, because ‘trying to imagine the disjunction is too complicated and distracting’.54 This seems true when we try to imagine the disjunctive reading that disjoins the five interpretations just discussed. It is not clear that the disjunctive reading of ‘The Ornamented Zither’ makes more sense than an interpretation that focuses simply on amorous love, music, marriage, political career, or poetry.
The second solution proposes to combine competing interpretations into a subsumptive reading of the following form:
W’s meaning is such that it is partly given by/aptly viewed under interpretation 1, partly given by/aptly viewed under interpretation 2, […] and partly given by/aptly viewed under interpretation n, where those embedded interpretations, I1, I2, and so on, are understood as first-order sub-interpretations subject to the higher-order interpretation I*, which subsumes them, though not simply disjunctively or conjunctively.55
Such an interpretation appears unwieldy and contrived. It is unwieldy because it is hard to describe what a work means with a simple phrase. ‘The Ornamented Zither’ becomes a poem partly about amorous love, partly about music, partly about official career, partly about poetry-making, and partly about the death of one’s spouse, where these interpretations are like the faces of a polyhedron that cannot be seen at one glance. The subsumptive approach is contrived because it seems easier to accommodate competing interpretations by adopting pluralism, where one can simply say that ‘The Ornamented Zither’ is multi-interpretable.
The third solution Levinson suggests is to introduce the aesthetic consideration as a tie breaker. That is, when we reach two or more epistemically best hypotheses, the one that makes the work artistically best is to be favoured.56 But this approach faces underdetermination again. I cannot see any sensible criterion forthcoming that can tell us which of the five hypotheses discussed earlier is artistically best. We have seen that the maximizer adopts a weakened notion of value-maximization precisely because of the problem of underdetermination. This being the case, the aesthetic criterion under consideration cannot be a real tie breaker.
IV. Concluding Remarks
Up to this point I have hoped to expose serious concerns with these theories by implementing an in-depth examination of Li Shang-yin’s ‘The Ornamented Zither’. Now, let us take stock of what has been argued so far. Anti-intentionalism is not workable because linguistic conventions are not sufficient to indicate a single right interpretation of a work. Additionally, the position alone fails to provide an adequate account of intentional use of language such as allusion and the like. The maximizing view weakens the force of anti-intentionalism and proposes to single out interpretations supported by textual evidence in terms of an aesthetic criterion. The aesthetic standard is construed as value-enhancing instead of value-maximizing. Nonetheless, this weakening of value requirement renders the requirement itself superfluous, since interpretations supported by textual evidence are bound to be value-enhancing in one way or another.
Actual intentionalism falls short of a useful theory in the typical case of interpretation, because in such a case extratextual evidence for the author’s intention is normally not available. Nor is the text the best evidence for the author’s intention; if it were, interpretative disputes would be easily resolved. The best the actual intentionalist can do in the typical case of interpretation is to have a bet on the author’s intention, which makes her a hypothetical intentionalist in practice. But hypothetical intentionalism faces the problem of underdetermination because, in the absence of the author’s intention, evidence always underdetermines the best hypothesis. The remedies offered by the hypothetical intentionalist are either unpromising or problematic.
Based on these reflections, I conclude that all of the four mainstream theories of interpretation have serious problems to be addressed by their proponents. If supporters of these theories want their accounts to be sustained in the debate over interpretation, they will need to address the worries I raise.
Notes
[1] Noël Carroll, ‘Interpretation’, in Routledge Companion to Philosophy of Literature, ed. Noël Carroll and James Gibson (New York: Routledge, 2016), 302–12.
[2] In my discussion, I will follow the Chinese name convention of presenting the family name first, since this is the convention adopted in the relevant literature cited in this paper.
[3] Hans Maes, ‘Intention, Interpretation, and Contemporary Visual Art’, British Journal of Aesthetics 50 (2010): 124–30.
[4] William K. Wimsatt and Monroe C. Beardsley, ‘The Intentional Fallacy’, Sewanee Review 54 (1946): 468–88.
[5] James J. Y. Liu, The Poetry of Li Shang-yin: Ninth-Century Baroque Chinese Poet (Chicago: University of Chicago Press, 1969), 13. This book is probably the most authoritative study of Li Shang-yin available in English up to now.
[6] James J. Y. Liu, ‘Li Shang-yin’s Poem “The Ornamented Zither”’, Journal of the American Oriental Society 85 (1965): 129–30.
[7] I say ‘at least’ because line one can be associated with another allusion, which will be mentioned later.
[8] ‘Chuang Tzu’ is nowadays more commonly translated as ‘Zhuangzi’. I will adopt this more common translation in my discussion.
[9] I cannot elaborate every interpretation owing to space constraints. See Liu, ‘Li Shang-yin’s Poem’ for details.
[12] This legend is recorded in Records of the Grand Historian, and Liu offers a different interpretation. See Liu, ‘Li Shang-yin’s Poem’, 130.
[13] Robert Ashmore, ‘Recent-Style Shi Poetry: Heptasyllabic Regulated Verse (Qiyan Lüshi)’, in How to Read Chinese Poetry: A Guided Anthology, ed. Zong-qi Cai (New York: Columbia University Press, 2008), 197.
[16] Wimsatt and Beardsley, ‘Intentional Fallacy’; Monroe C. Beardsley, The Possibility of Criticism (Detroit, MI: Wayne State University Press, 1970).
[19] Monroe Beardsley, Aesthetics: Problems in the Philosophy of Criticism, 2nd ed. (Indianapolis: Hackett, 1981): 122–26.
[23] Michael Wreen, ‘Beardsley’s Aesthetics’, in Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta and Uri Nodelman, spring 2023 ed. (Stanford University, 1997–), https://plato.stanford.edu/archives/spr2023/entries/beardsley-aesthetics.
[24] David Davies, Aesthetics and Literature (London: Continuum, 2007), 73.
[25] Jerrold Levinson, The Pleasures of Aesthetics: Philosophical Essays (Ithaca, NY: Cornell University Press, 1996), 177.
[29] Sherri Irvin, ‘Authors, Intentions and Literary Meaning’, Philosophy Compass 1 (2006): 120–22.
[30] Stephen Davies, Philosophical Perspectives on Art (Oxford: Oxford University Press, 2007), 166–90; Alan H. Goldman, Philosophy and the Novel (Oxford: Oxford University Press, 2013).
[31] I am using ‘aesthetic value’ in the broad sense to include what some philosophers call ‘artistic value’.
[32] Robert Stecker, Aesthetics and the Philosophy of Art: An Introduction (Lanham, MD: Rowman & Littlefield, 2010), 147.
[34] Berys Gaut, ‘Morality and Art’, in A Companion to Aesthetics, ed. Stephen Davies et al. (Malden, MA: Blackwell, 2009), 429.
[36] Eric D. Hirsch, Validity in Interpretation (New Haven, CT: Yale University Press, 1967); Steven Knapp and Walter B. Michaels, ‘Against Theory’, Critical Inquiry 8 (1982): 790–800; Noël Carroll, ‘Art, Intentions, and Conversations’, in Intention and Interpretation, ed. Gary Iseminger (Philadelphia, PA: Temple University Press, 1992), 97–131; Robert Stecker, Interpretation and Construction: Art, Speech, and the Law (Oxford: Blackwell, 2003); Paisley Livingston, Art and Intention: A Philosophical Study (Oxford: Oxford University Press, 2005); Maes, ‘Intention’; Kathleen Stock, Only Imagine: Fiction, Interpretation and Imagination (Oxford: Oxford University Press, 2017).
[39] Noël Carroll, ‘Interpretation and Intention: The Debate between Hypothetical and Actual Intentionalism’, Metaphilosophy 31 (2000): 77; ‘Art Interpretation’, British Journal of Aesthetics 51 (2011): 128; ‘Interpretation’, 308.
[47] Kendall L. Walton, ‘Categories of Art’, Philosophical Review 79 (1970): 334–67.
[53] Gregory Currie, Arts and Minds (Oxford: Oxford University Press, 2004), 131–32.
[55] Jerrold Levinson, Contemplating Art: Essays in Aesthetics (Oxford: Oxford University Press, 2006), 304–5.
Acknowledgements
My sincere thanks to Stephen Davies, Fred Kroon, Cynthia Freeland, and the reviewers for this journal for their assistance with this paper.
Funding Information
This work was supported by Taiwan’s Ministry of Science and Technology (now the National Science and Technology Council). Project number: 108-2410-H-031-085-MY3.
Competing Interests
The author has no competing interests to declare.