7. E[ric] D[onald]
Hirsch1
[1. The key for Hirsch citations is: AIM: Aims in
Interpretation (1976); IE: Innocence and Experience (1964); PC:
“Privileged Criteria in Literature” (1969); and VAL: Validity in
Interpretation (1967).]
Don Hirsch entered the field of
"general hermeneutic theory" in order to confront “the problem of
validity," which “has been neglected in recent years” (VAL viii). He
avowed that "valid interpretation is crucial to the validity of all
subsequent inferences” in “all human studies.” “The theoretical aim of a
genuine discipline, 2 [. Hirsch's critical vocabulary is
loaded with terms that imply a judgment of what is authentic (e.g.
“genuine," “serious,"). The actual state of literary studies
appears spurious; the real is made to seem unreal.], scientific or humanistic,
is the attainment of truth, and its practical aim is the agreement that truth
has probably been achieved” (VAL viiif). “The practical goal" is
therefore the “consensus" that “one set of conclusions is more probable
than others" (VAL ix). Though “the subject matter of interpretation is
often ambiguous and its conclusions uncertain," “valid interpretation”
is needed for “any humanistic discipline to claim genuine knowledge” (VAL
viiif, e.d.). Hirsch wants to reafflirm the “intellectual respectability” of
"the activity of interpretation” by demonstrating how “its results can
lay claim to validity” (VAL 164).
Hirsch envisions nothing less
than a “general hermeneutics” “for all textual interpretation” (AIM 17).3
[3. On the history of the controversy over hermeneutics in science versus
humanities, see AIM 150f. The fact that hermeneutics was originally practised on
sacred texts whose authors claimed divine authority is interesting in light of
Hirsch's exaltation of authorial meaning and intention.] To that end, he denies
we have any "reason for isolating literature and art” from “other
cultural realities" (AIM 109). “The literary text has no special
ontological status" (VAL 210).4 [4. Hirsch's suggestion
that any claims for the special status of literature have a “mystical"
nature (VAL 210; AIM 109) is quite unfair to stylistic and linguistic theories,
which he spurns precisely because they are too deterministic (c£ AIM 50f). His
project depends crucially on this thesis and could succeed only to the extent
that we can set aside the special functions of literature described by most of
the critics I review, particularly Wellek and Warren, Frye, Fiedler, Iser, jauss,
and the Yale group.
For related motives, Hirsch judges
the traditional opposition between the “hard” or “exact” sciences versus
the “soft” or “inexact” humanities to be a "nonsequitur" (AIM
149). "The cognitive elements in both have exactly the same character": 'the progress of knowledge and its
consolidation are governed by the critical testing of hypotheses with reference
to evidence and logic" (AIM 149, 151). 'Knowledge in all fields thus turns
out to he a process rather than a static system" (AIM 152). Every
“discipline”5 [5. 'Discipline' is intended to render the
German term "Wissenschaft" applied to both the sciences and
humanities, though the “equivalent is not close enough” (AIM 150). is “'a
communal enterprise” for whom “the logical relationship between evidence,
hypothesis, and probability” "remains the paradigm (or
ideology!)."The 'enterprise" must make sure 'past evidence is
stored" and “unfavourable evidence” is not “suppressed," but
brought to bear “upon a hypothesis to which it is relevant.” “The process
of knowledge ceases” when “the consolidation and discovery of evidence
decline”' along with “the commitment to the critical testing of hypotheses
against all known relevant evidence” (AIM 153).
This portrait of “knowledge" as a "process" is widely
shared, even by the subjectivist Bleich. 6 [6. Despite their
diametrically opposed philosophies, several parallels between Hirsch and Bleich
can be found. They both advocate negotiation of individual meanings in order to
attain consensus. Both hold that the text exists only when it is read. Both
consider subjectivity the norm in reading. Both draw on Piagetian psychology.
And both show great respect for the author, though this move fits Bleich's
scheme rather badly (see Ch. 11).] The performative nature of cognition is, as 1
argued in Ch. 1, steadily gaining recognition in many disciplines. Hirsch
proposes to resolve the subject-object division with the thesis that
“objectivity consists in the universality of the subjective experience” (AIM
99). But his whole method reveals the anxiety that no such "universality”
prevails, so that it must be imposed by painstaking procedures for. choosing and
rejecting. The outcome would tend to drive a still greater wedge in between
subject and object by objectifying various stages and results in the process of
knowledge so as to blot out the issues of subjectivity: bias. selectivity,
motivation, disposition, interest, and so on. These issues are extremely
relevant for many uses of literature in society, and the reasons for severing
them from interpretation need to he justified.
Hirsch rests his hopes on logical procedures that attain “validity”
by “relating hypotheses” to a “body of evidence" (AIM 151) and
deciding which meaning is the most probable. The concept of “probability"
applied here is derived not from the "statistics' of 'numerical
quantities,' but from "the logic of uncertainty” (VAL 173f). The main
tactic is to narrow down the “classes” to which a disputable “object"
belongs (VAL 176ff). Since 'the idea of the class in itself entails an idea of
uniformity” based on “the defining characteristics of the class,"
“'anything we can do to narrow the class" will make the characteristics
more specific and “increase the likelihood” of our “judgment" being
“true” (VAL 176, 179). For instance, if a word-meaning is in dispute, we
could work through classes like these: (a) uses of the word in English at large;
(h) uses in the historical period of the text; (c) uses in both historical
period and genre of the text; (d) uses by that same author in that genre; (e)
uses in that same text; and (f) uses in that same passage (cf. VAL 184ff). However, neither disputes nor
evidence are usually this clear-cut.
Somewhat paradoxically, the
narrowing of classes is to be achieved by expanding one's materials. The
interpreter “should base his decision on all the relevant evidence
available" in order to make “a grounded choice between two disparate
probability judgments on the basis of common evidence which supports them”
(VAL ix, 180). Hirsch concedes “it would be unfeasible and undesirable to
publicise all the evidence relative to every interpretive problem” (VAL x).
However, he conjures literary scholars to “take the responsibility of
adjudicating the issue in light of all that is known”; the fact “that few
such adjudications exist merely argues strongly that many more should be
undertaken” (VAL 171). Hirsch himself, though, has not “undertaken” any
(as he freely admitted to me), for reasons which may become clearer in the
course of my review.
Repudiating the objectifying notion
that the meaning is “in” the text, Hirsch acknowledges that “the text does
not exist even as a sequence of words until it is construed” (VAL 13) (this
too a thesis of Bleich's). But Hirsch encounters a dilemma between two wholly
disparate notions about how that “construing” takes place. On the one hand,
he wants to make the processes of comprehension look disorderly enough that his
stringent methods of interpretation will seem vitally necessary, like a
deliverance in an hour of dire need. On the other hand, he wants to suggest that
those processes are elaborately controlled, because his methods require
determinate and stable meanings to work on.
In the 1967 volume (VAL), the
disorderly version gets prominent coverage. In contrast to many theoretical
critics, Hirsch declares that “the process and psychology of understanding are
not reducible to a systematic structure" (VAL 170). This judgment is
reached by applying a rigorous standard: “there is no way of compelling a
right guess by means of rules and principles.” He observes “there would not
be any problem of interpretation” if “public unanimity" “existed
generally” about the “meaning of a text” (VAL 120. “The variability of
possible implications is the very fact that requires a theory of interpretation
and validity” (VAL 123).
Hirsch follows Schleiermacher (1959
[1838]: 109) in dividing literary response into two stages: a “divinatory
moment” that is “unmethodical, intuitive, sympathetic"; and a
“critical moment” that “submits the first moment” to a “high
intellectual standard” by “testing it against all the relevant knowledge
available” (VAL X).7 [7. Schleiermacher quaintly labelled
these “functions” “female” for the “divinatory” (“ transforming”
oneself “into the author”) and “male" for the “comparative”
(moving from “general” to “unique”) (VAL 204). Mercifully, Hirsch does
not adept this terminology.] These two “correspond to two distinct moments in
knowledge which Whitehead aptly calls “the stage of romance” and “the
stage of precision." Hirsch's theory “is mainly concerned with the second
moment,” the only one to "raise interpretive guesses to the level of
knowledge." The first moment resembles “a genial guess” and is not
determined by validity: "there are no methods for making guesses, no rules
for generating insights" (VAL 204). “The methodical activity of
interpretation commences when we begin to test and criticize our guesses.”
This proposal resolves Hirsch's
dilemma by splitting reading into a disorderly initial phase and a totally
ordered final phase. He grants that “these two sides of the interpretive
process, the hypothetical and the critical, are not of course neatly separated
when we are pondering a text, for we are constantly testing our guesses both
large and small as we gradually build up a coherent structure of meaning"
(VAL 203f). Yet his theory is expressly justified with the purported opposition
between “the whimsical lawlessness of guessing" and the “ultimately
methodical character of testing”: “both processes are necessary in
interpretation, but only one of them is governed by logical principles."
In the 1976 volume (AIM),
Hirsch attenuates the split: “the private processes of verbal understanding
have the same character" as the “public activity” of .validation,”
the “objective marshalling of evidence in the cause of an interpretive
hypothesis” (AIM 33). Hence, “the process of validation is not easily
separated from the process of understanding in either theory or practice.” He
does not think this change of perspective demands "substantive revisions of
the earlier argument” (AIM 8). Yet if his two stages of reading have "the
same character,” his elaborate enterprise of validation, advocated on the
grounds of their difference, seems far less compelling. He also owns that he had
previously "almost" “ignored the whole question of the process of
understanding” (AIM 33). By dividing reading into two “moments" and
offering a theory only for the second, he essentially proposed a theory that
deals with effects, yet pays no attention to causes. It seems illogical to
classify and judge competing hypotheses while disregarding how they are
engendered in the first place. That way, we can only objectify them as entities
which abruptly emerge ready-made from a spree of “lawless guessing."
Hirsch remarks that “meaning" “is not a physical object," and to
treat "meaning as an object” is merely “a short-cut” or
“convenience" to designate both “intentional objects” and
“intentional acts” (AIM, 8). But this short-cut conceals precisely what most
theoreticians on the current scene consider the vital issue: how “objects of
knowledge” or “objects of our construing" (VAL 176f) are created by the
human subject participating in literary communication.
Hirsch stipulates that personal
interests in formulating hypotheses should be completely discounted. He finds it
“obvious" that "the consolidated knowledge
within a
discipline has nothing to do with its rhetoric” (AIM 153). Through its means
of "persuasion," “rhetoric” influences “the communal acceptance
of hypotheses” and “can subserve both knowledge and intellectual chicancry.8
[8. Perhaps the expansion of rhetoric in the theorizing of the Yale critics (Ch.
14-16) is one motive why Hirsch attacks the “American disciples" of
"Derrida and Foucault' (AIM 147,13). His campaign against these
“cognitive atheists” will be described later.] Although "the spirit of
vanity and advocacy” always endangers the “selfless devotion to the communal
enterprise,” “the direction of knowledge goes forward at the level of the
discipline” (AIM 152f). It is “essential to distinguish hypotheses from the
rhetoric used to convey them,” since they "are not bound to any single
expression of them” (AIM 153f). This demand might be met by stating every
hypothesis in several forms, then proving them partially equivalent, and finally
extracting only what is common to them all. This task looks intriguing, but we
do not find it performed in Hirsch's books beyond brief demonstrations for a few
parts of texts.
Given
the arguments summarized above, we would not expect Hirsch's orderly version of
reading to be depicted in psychological terms. Nor can it be a product of
“linguistic norms,” since they constitute “the possibilities, not the
actualities of language,” and may therefore “be invoked to support any
verbally possible meaning" (VAL 69, 226). Instead, he proposes an
essentially set-theoretical concept of the “intrinsic genre."9
[9. Saussure's stale dichotomy of “langue versus parole” is invoked, but the
“intrinsic genre” is not to be “subsumed under either category” (VAL 69,
111).] It seems to emerge during a movement from the former to the latter. as an
“overarching notion” which “embraces a system of expectations” and
“conventions,” including “the entire system of usage traits, rules,
customs, formal necessities, and properties which constitute a type of verbal
meaning” (VAL 78, 92). If "genre" is defined as the “type which
embraces the whole meaning of an utterance" and “controls the temporal
sequence of speech,” then "all understanding of verbal meaning is
necessarily genre-bound” (VAL 71, 78, 76). “The intrinsic genre of the
utterance” is thus “the essential component of a context”; “everything
else in the context serves merely as a clue to the intrinsic genre” (VAL 87).
The “shared genre conception” is “constitutive both of meaning and of
understanding" (VAL 80f). These theses lead to a "universally
applicable" “principle”: “valid interpretation depends on a valid
inference about the properties of the intrinsic genre” (VAL 121).
Hirsch stresses that his idea
of “genre" is far more specific than traditional literary genres like
“Christian-humanist epic”; indeed, “there is no ready-made vocabulary for
describing the intrinsic genres of particular utterances" (VAL 84, 82).
Hirsch is willing to admit the old "broad genre concept,” as long as it
does not “pretend to be a species concept that somehow defines and equates the
members it subsumes" (VAL 110). But he rejects “the notion that the
larger classifications of texts represent an adequate foundation for defining
different types of interpretation” (VAL 113). “There are no clear and firm
boundaries" among those “classifications,” and their
“categories" are "not everywhere equally " (VAL 115ff). The
“broad genre concept" tends to rely on “conclusions about recurrent
patterns” "subsequent to interpretation," as opposed to the
intrinsic genres actively applied for and during interpretation (VAL 110).
We can see that even on the
theoretical plane, Hirsch is already practising his precept of narrowing down
classes (in this case, "genres”). In principle, he “objects to the
dangerous practice of using abstract categories or monolithic approaches” and
“methods” to interpret a wide variety of texts” (VAL 88f). He feels that a
“description” of “the common elements in a narrow group of texts which
have direct historical relationships” "becomes less useful to
interpretation” as its "scope becomes broader and more abstract"
(VAL 110).
The "intrinsic
genre," in contrast, “emerges” during communication "only after a
narrowing process," going from “vague and empty" toward “more
explicit" (VAL 103, 77). Indeed, his notion of "genre" appears so
specific that Hirsch reassures us it is not “identical with the particular
meaning of the utterance" (VAL 86). “An interpreter's preliminary generic
conception of a text is constitutive of everything else that he subsequently
understands" and provides an “anticipated sense of the whole by virtue of
which the presently experienced words are understood in their capacity as parts
of a whole" (VAL 74, 82). Therefore, the "intrinsic genre" is
definable as the "sense of the whole by means of which an interpreter can
correctly understand any part in its determinacy" (VAL 86) -- precisely the
construct needed to support a “valid interpretation.”
The “intrinsic genre"
also “determines” "the implications of an utterance,” a “crucial
issue" “when our central concern is validity" (VAL 89). The genre is
not merely the “additive” “set of implications" postulated in logic,
but a means for “structuring" and "unity,” and an indicator of
“purposes,” "emphases," and “relative importance" (VAL 98f,
117, 102). Control is thereby maintained without total determinism. “Minor
alterations can always be made” in the text “without changing the intrinsic
genre” (VAL 85). Conversely, the same "word sequence can represent more
than one meaning” because it “can be subsumed by more than one intrinsic
genre and therefore can carry different implications" (VAL 98). It follows
that “the correct determination of implications" is “crucial" for
“discriminating a valid from an invalid interpretation" (VAL 89). By the
same token, "disagreement about an interpretation is usually a disagreement
about genre" and “centres on details of implication” (VAL 98, 89).
This conception of a
“genre" yielding a “sense of the whole” is disrupted somewhat when
Hirsch disclaims “coherence” as a measure for validation. His reasoning
follows logic: “appeals to coherence” are "useless because they are
circular” (VAL 194, 237). Though “the hypothesis which makes functional the
greater number of traits must, in relation to that limited evidence, be judged
the more probable,” this adjudication is “unsatisfactory" because
“one hypothesis will make functional different traits from the other"
(VAL 190). Hirsch wants the interpreter to “consider all the known relevant
data” (VAL 192), as if relevance itself could be better decided by any other
criterion than the coherence among the items it is assigned to – a notion
which I hold, on the basis of the vast evidence amassed in my own books, to be
nonsense.
In the orderly version,
reading generates meanings that, in Hirsch's view, are always determinate,
because “determinacy is a necessary attribute of any sharable meaning"
(VAL 44). This conception is supported with the logical principle of identity:
“if a meaning were indeterminate," “it would have no boundaries, no
self-identity, and therefore could have no identity with a meaning entertained
by someone else. “ “Verbal meaning” "is what it is and not something
else, and it is always the same” (VAL 450. “Determinacy does not mean
definiteness or precision” or “clarity" (VAL 44, 85). As we see, Hirsch
postulates inevitable determinacy by construing the concept in a very weak
sense, as compared to the everyday sense of “definitely settled,"
“conclusively determined,” “unequivocally characterized” (Webster's
New Collegiate Dictionary).
Logicians have yet to
prove that the principle of identity applies as straightforwardly to meanings as
to physical objects, and I do not believe they ever will. Hirsch makes a foray
in this direction when he defends “the existence and importance of synonymity
[sic], that is, the expression of an absolutely identical meaning through
different linguistic forms” (AIM 50). He argues that “form does not compel
meaning," so that “the relationship" between the two is
“essentially indeterminate" (AIM 10, 50f). He opposes “linguistic
determinism" with something he styles the “Gödel's theorem of
language": “the intrinsic undecidability of the correlations between
linguistic levels” (AIM 50f, 66; cf. Bazell 1953, 1966). This attitude might
explain why Hirsch rejects stylistics as a "reliable method of confirming
an interpretation” (AIM 22, 50, 72).
Synonymity also is construed in
a weak sense. For Hirsch, it requires that expressions can be “substituted”
for each other only “occasionally” rather than "universally” (cf. AIM
54). Expressions may appear different in isolation, but equivalent in a
particular text (cf. AIM 61). 10 [10. He cites a “test” or
“experiment” he did on synonymity by “showing” “documents” “to a
number of literate native speakers” (AIM 60f), but does not spell out method
or results). This proviso removes the obstacles that “language" allows
“the same words different meanings” and “different words the same
meaning'; and that “meaning postulates” are “only provisional,” subject
to change during “actual use" (AIM 620. Even poetry gets included by
reasoning that we “cannot reliably assert” “absolute synonymity is
impossible” in “poetry" “unless we can also assert non-synonymy for
all speech" (AIM 58f). This reasoning depends on Hirsch's original denial
of a “special ontological status” to “the literary text" (VAL 210).
No matter what a dispute may
concern, a scholar cannot compare objects and classes without first constituting
them. This process cannot be taken for granted, and Hirsch provides no
directions for performing it upon complex materials. Empirical studies reveal
that humans have only fuzzy notions of the criteria for defining even familiar,
concrete classes of objects; and that class membership is not uniform, but
arrayed along a gradation of more or less typical (cf. Rosch & Mervis 1975;
Rosch 1977). If so, the far less ordinary or tangible objects of creative
literature could hardly fall into orderly and homogeneous classes. Hirsch would
restrict innovation to “novel subsumptions under previously known types"
(VAL 105). 11 [11. The possibility of “unique meaning” is
denied because “understanding” depends on knowing what “type of meaning”
to expect (VAL 80). Wellek and Warren (TL 18, 151) offer a similar argument.
Iser suggests on the contrary that each experienced meaning is
“unrepeatable” (AR 150).] Yet he elsewhere depicts literature as "an
arbitrary classification of linguistic works which do not exhibit common
distinctive traits” (AIM 135). So this domain should be the least amenable one
to logical classification based on the “uniformity” of “defining
characteristics" (VAL 176).
To forestall such problems,
Hirsch further constricts and stabilizes the objects of “interpretation.
" He limits them to what he calls “verbal meaning,” as opposed to a
whole gallery of meanings that would disrupt his “validation:
“significance,” the “relationship between that meaning and a
person, or a conception, or a situation, or indeed anything imaginable";
“subject matter;” "the “objective” character" of what a
“verbal meaning" “refers to"; "response," “the more or
less personal meaning” a reader "attaches to a verbal meaning";
“unconscious and symptomatic meaning,” "involuntary accompaniments to
meaning"; and “implication," i.e. "any sub-meaning belonging to
the whole array" of meanings .carried” by “an utterance” in the same
way that “a trait belongs to a type" (VAL 8, 59, 39, 52f, 62, 71). The
(wholly gratuitous) problem now is that “we cannot isolate the act of
construing verbal meaning from all those other acts, perceptions, associations,
and judgments that accompany that act and are instrumental in leading us to
perform it" (VAL 140). “Such aspects of a context as purposes,
conventions, and relationship to an audience are not outside the meaning of the
utterance but are constitutive of it” (VAL 87). Besides, we already heard how
Hirsch stressed the “crucial" role of “implications” for telling what
is “valid".
A skilled interpreter might try
to sort out different kinds of meaning after having already arrived at a
provisional interpretation. But quite apart from the difficulty of attaining
such specialized skills, the interpreter would be in the odd position of
discarding for theoretical motives some elements of meaning that he or she had
depended on in practice, rather like a magician disclaiming the mechanisms
behind the act. If the purported “verbal meaning" could not have been
attained without using these other types of meaning, it can hardly have a
sufficiently independent status to provide the structural base for an elaborated
process of validation. Even if subjectivity can be trimmed down close to
objectivity, subsequent inferences are not automatically objective.
Nonetheless, Hirsch draws a
line between "interpretation, whose exclusive object is verbal
meaning," and “criticism,” whose “proper object” is the text's
.significance" -- a domain he considers “boundless,” invested with
“countless dimensions” (VAL 57). Though he says “all textual commentary is
a mixture of interpretation and criticism," he thinks that "usually a
choice has been made as to which goal is to receive the main emphasis” (VAL
140). He advertises this “distinction” as a “charter of freedom to the
critic, not an inhibition” (VAL 57). But this freedom is rescinded again when
Hirsch declares that "the discipline of interpretation is the foundation of
all valid criticism" (VAL 156). Here we have the typical move of the critic
portraying his own theory as central and indispensable. The implication should
be that a large portion of available criticism is "invalid" because it
doesn't conform to Hirsch's methods. But in my discussions with him, he has
repeatedly claimed to be merely providing a theoretical rationale for what
interpreters have been doing all along.
The problem of ambiguity is
treated not as flux or dispersal, but merely as an array of alternate
determinate meanings awaiting a decision. "Whenever a reader confronts two
interpretations which impose different emphases on similar meaning
components,” he says, “at least one of the interpretations must be
wrong" (VAL 230). This assertion reverts back even before the work of the
New Critics, notably Empson's (1930) classic study showing ambiguity being
systematically present in literary communication without pressure to decide
which readings are “wrong.” Hirsch imagines that Empson's argument depends
on the text being “conceived" as "a “piece of language" (VAL
62, 224). But interpretive hypotheses must become such pieces too before they
can be adjudicated; and how to make sure they convey only “verbal meaning”
without "significance" is far from obvious.
Whether the orderly or the disorderly
version of reading is accepted turns out to be inconclusive, because Hirsch
situates the ultimate recourse not at the reader's end, but at the author's. The
"philological effort to find out what the author meant” is “the only
proper foundation of criticism" (VAL 57). “The only object of cognition
having an implicit claim to be “a universal point of reference” is the
sharable object cognized by its maker” (AIM 105f). "All valid
interpretation of every sort is founded on the re-cognition of what an author
meant" (VAL 126). The term “re-cognitive" was adopted from Emiliano
Betti (1955), who contrasts it with “presentational” and “normative”
interpretation (VAL 112). Though Hirsch asserts that “only a re-cognitive
interpretation is a valid interpretation" (VAL 122), his own standards are
heavily normative, at least in the sense that “the normative dimension of
interpretation is always in the last analysis an ethical dimension” (AIM 77).
But the ethical basis of his enterprise was not foregrounded in his earlier
formulations. I shall return to this shift later.
As we might predict, Hirsch
finds the decisive act is not so much in speaking or writing, but in selecting
an “intrinsic genre.” “Once the speaker has willed “this particular type
of meaning,” the further determination of his meaning depends entirely upon
his subsequent choice of words and patterns falling within the tolerance of the
intrinsic genre” (VAL 86). Hirsch presumes that the author's original act of
"will" to convey a "type” (VAL 31, 51, 86, 124, etc.) always
has a specific result. We are categorically told that “an author cannot mean
what he does not mean”; that "no example of the author's ignorance with
respect to his meaning could legitimately show that his intended meaning and the
meaning of his text are two different things"; and that "either the
text represents the author's verbal meaning or it represents no determinate
verbal meaning at all" (VAL 22, 234). These are interesting claims about
psychological states, although Hirsch offers no empirical evidence to support
them, and I doubt any could be found.
He sidesteps the “distinction
between a mere intention to do something and the concrete accomplishment of that
intention” by transferring it from the domain of “verbal meaning" over
to that of “evaluation" (VAL 11f). There too, the author is the point of
reference. “The only values which can be considered intrinsic properties of a
work are those which attach by subjective necessity to a re-cognition of the
author's work" (AIM 106). In contrast to Frye, Hirsch supposes that
“evaluation” can "qualify as objective knowledge” if it is
"accurate with respect to" “explicit” “criteria being
applied" (AIM 108). Such “judgments” “furnish the grounds of their
own validation" and can attain “as much objectivity as accurate
interpretations.” Reciprocally, "those which are necessarily implied in
interpretation” are “the only unavoidable judgments of value in literary
commentary" (AIM 106,). "it is quite possible to eschew other
kinds.”
Such a strong stand for the
author is highly atypical, and we should consider the difficulties (beyond mere
fashions of the day) that make it so unpopular. One difficulty is the sheer size
of the "philological effort” involved. In order to find out what the
author meant, we are enjoined to consult “all clues" about the
"cultural and personal attitudes the author might be expected to bring to
bear in specifying his verbal meanings” (VAL 240). Interpreters might have to
sift through a staggering mass of historical, biographical, and psychological
evidence on the good faith that it will apply to the issues being adjudicated.
But the assembled evidence might be inconclusive, contradictory, or unrelated to
the text or passage in dispute. Besides, we have no reliable way to tell exactly
when .all” the evidence is in. For some authors, an entirely scholarly career
might be spent digging and gathering without producing the necessary certainty.
A second difficulty is the
incongruity of trying to extract the narrowest possible meaning from the widest
possible scope of evidence. To make meaning decidable, we are asked to make it
undelimitable. Hirsch agrees with Wellek and Warren's “programmatic idea”
that “literary interpretation must be intrinsic” (VAL 113). But the mass of
“clues" are highly likely to spill far outside the boundaries of the
clearly determinate “verbal meaning” Hirsch would make the sole basis for
interpretation. Such clues could easily belong to “significance,”
“implication,” and all the other "meanings” we saw Hirsch setting
aside; even “autobiographical meanings” are discounted (VAL 16). Therefore,
the ostensibly “objective" evidence only becomes evidence at all when the
interpreter subjectively elects to consider it as such. Hirsch himself admits
that since “the intrinsic genre is always construed, that is, guessed, and is
never" "given,” “the interpreter can never be completely
certain" what “genre" the speaker has "willed" (VAL 88,
94).12 [12 Probably influenced by ordinary-language
philosophy, Hirsch uses the term “speaker” in much of his argument and
alludes to Saussure's obtuse view that “writing is a lately developed
surrogate of speech” (VAL 101f).. But in fact, writing poses problems of
interpretation that speech does not, or in different ways (cf. survey and
references in Beaugrande 1984a). For one thing, getting validation from a
speaker should be easy.] “We have no direct access to the author's mind”
(VAL 99). "The speaker's attitudes are not given but are construed from the
utterance itself” (VAL 87). The investigator often has to rely on an
“arbitrary supposition” derived from a “psychological reconstruction” of
the author (VAL 123, 240). We seem to be trading a community of unreliable
readers for the single authority of an inaccessible author.
A third difficulty is that text
production is being pictured as stable and reliable, even though text reception
has been portrayed as “whimsical” and “lawless” (VAL 203f). There is no
empirical support for the idea that when people turn from reading to writing,
their mental processes abruptly become more determinate. Available findings
indicate rather that authors typically work from a highly approximative mental
representation of intended meanings and, for a whole spectrum of reasons, may
fail to execute their intentions. 13 [13 On the processes and
problems of writing, see Beaugrande (1984a). The failure of authorial intention
becomes thematic in the theorizings of de Man, Bloom, and Hartman (Ch. 14-16).
Compare also Notes 8 and 16.] We
might therefore incur the additional labour of proving that whatever authors say
about their meanings -- even explicit paraphrases, which would presumably
constitute the most directly applicable evidence -- is indeed dependable. The
consensus among our critics (notably, Fiedler, Iser, Holland, Paris, and de Man)
is more the contrary. Even Hirsch admits that in literature, the author
“submits to the convention that his willed implications must go far beyond
what he explicitly knows” (VAL 123).
A fourth difficulty is that the
author's intention might be precisely to evade determinacy. As several of our
critics point out, Iser in particular, modernism is distinguished by a split or
circuitous relation between expression and intention, a deliberate withholding
of clues for readers who wonder where the author's real opinions lie. It is no
accident that none of Hirsch's samples are taken from modernist texts, a problem
I shall return to. In effect, we are urged to revere all authorial intentions
except those which prove inconvenient for the project of validation. And this
injunction too seems illogical, or, from an ethical standpoint, inconsistent.
Hirsch's theory implies the
general hypothesis that the literary author uniformly intends to convey a single
"valid" “verbal meaning,” no matter how much interpretive effort
might be required to establish it. This assumption looks dubious when we
consider the overwhelming diversity of intentions that lead authors to compose a
work. If we juxtapose Homer with Milton, Richardson with Fielding, Goethe with
Sterne, Zola with Kafka, Wordsworth with Berryman, Twain with Joyce, or
Hauptmann with lonesco, we may seriously doubt any uniformity of intentions
regarding determinate and decidable meaning. Instead of just amassing clues
about “verbal meanings,” we might seek clues about whether authors did or
could have intended standards of "validity" in the first place. Like
Wellek and Warren, Hirsch offers us only his coercive warning that the
alternative to his own view is mere anarchy -- the absurd doctrine that an
author or text “does not mean anything in particular" (cf. VAL 4, 11, 13,
45, 234).
Hirsch is sufficiently aware of
the problematic character of his basic assumptions to have anticipated some
possible objections to them. In his earlier book, he professes to refute his
adversaries with logical argumentation. He charges them with various forms of
“scepticism” that “implicitly deny the possibility of validity in any
absolute or normative sense of the word" (VAL viii). His refutations rest
mainly on the same tendencies we have seen in his whole model: splitting of
subject from object, and adducing relative probabilities.
In one form of “scepticism”
called “psychologism," “textual meaning" "changes from
reading to reading" (VAL 6). He retorts that it is a “mistake” to
“identify meaning with mental processes rather than with the object of those
processes" (VAL 32). “The objects of awareness are not the same as the
subjective “perceptions,” “processes,” or “acts” which are directed
toward those objects” (VAL 37). If .an unlimited number of intentional acts
can intend (he averted to) the very same intentional object,” they can also
“intend the same verbal meaning" (VAL 38). Thus, it is “possible to
reproduce a verbal meaning,” even though “one man's mental life is not the
same as another's” (VAL 38, 32). “It is far more likely that an author and
an interpreter can entertain identical meanings than that they cannot” (VAL
18). Of course, it is narrow “verbal meaning" rather than wide
“significance" that is “in principle reproducible” and
“sharable" (VAL 38, 40).
In a second form of
“scepticism” called “radical historicism," “the meaning of the
literary text is “what it means to us today" (VAL viii). Hirsch replies
that a “reinterpretation is not the same as a different understanding"
(VAL 42). After all, “all understanding of cultures past or present is
“constructed"; “there is no immediacy in understanding either a
contemporary or a predecessor” (VAL 43).14 [14. The
admittedly fanciful attempts of Roland Barthes and Jan Kott to make Racine and
Shakespeare into our contemporaries are cited to prove that the contemporary
mind is always a “construction” (AIM 41, 88).] He opines that “generally,
we are more likely to get a contemporary text right” than “a text from the
past”; “but this general likelihood does not automatically hold in any
particular instance (where factors of temperament, knowledge, diligence, and
luck are decisive)” (VAL 44). This argument plays upon the “distinction”
between “the general probability” versus “the particular probability that
may obtain in a particular case” (VAL 42f). Apparently, the most probable
meaning is to be uncovered by improbable acts of scholarly diligence.
In a third form of "scepticism” called “autonomism" the
"central tenet” is: “it does not matter what an author means -- only
what his text says” (VAL 10). This time, the rebuttal is that a text doesn't
“say” anything; there is only “the saying of the author or a reader”
(VAL 13). “Signs can be variously construed, and until they are construed, the
text “says nothing at all" (VAL 14). Although Hirsch presents the
“empirical fact" that “public consensus does not exist,” he
attributes this not to “private meanings," but only to
“improbable" or “wrong” ones (VAL 13ff). Also, the idea that
“meaning is independent of authorial will" would require the supposition
that “literary texts belong to a distinct ontological realm” (VAL viii) -- a
thesis Hirsch has to reject, as we have noticed (cf. VAL 210; AIM 58f).
In
more recent times, Hirsch adopts a different defense after transferring the main
rationale of his theory from the logical over to the ethical. He may have felt
angered by the refusal of criticism to take his proposals seriously. Or, he may
have realized that the latter entail an arduous task with an uncertain outcome.
Either way, he elects to meet any doubts about whether his “validation” is
being done or can he done by avowing with rising shrillness that it jolly well ought
to be done in the name of ethical responsibility. Even he can see how hard it is
to find explicit model interpretations that genuinely fulfill his criteria.
In the Validity volume, this lack of model interpretations was
acknowledged with equanimity. We “will be disappointed" if we “expect
to discover” there "a new interpretive program or approach" (VAL x).
Nor should we "expect to find complete and exemplary demonstrations of the
validating process." “The practical consequences” of such a theory
“are bound to he largely indirect"; he "believes” they will
“take care of themselves" (VAL xf).
In Aims, however, Hirsch finds they haven't. He is mightily
“indignant" to see how “the anxiety-ridden insistence on distinguishing
itself from natural science” has led “literary study” to become “the
most sceptical and decadent branch of humanistic study” (AIM 13, 149). The
“sceptics” of the earlier volume are restyled “cognitive atheists"
for “assuming that all “knowledge” is relative” (AIM 36). Though he
denounces his opponents as “theologians” (AIM 13; cf. VAL 44), his own term
“atheism” implies not a failure to he objective, but a failure to make a
leap of faith -- in this case, the faith in “validation." Theological
overtones pervade a project which began with an emphatic salute to scientific
method (just what Bleich would predict). Its “patron saint” is none other
than Matthew Arnold, who “was profoundly right to set the aims of criticism on
foundations that can ultimately be described in ethical terms” (AIM 139).15
[15. Hirsch is one of Arnold's few supporters in our survey. Wellek and Warren
complain that Arnold “confounded” “psychology" with
"evaluation” (TL 178). His rationale for measuring the greatness” of
works is pronounced “nonsense” by Frye (AC 22). His "elitism” is
assailed by Fiedler (WL 104, 128) and Bleich (AX 111f). His “intellectual
position" is transforrned by Holland into an “unconscious” “wish to
avoid sexual touchings” (DY 156). Hartrnan laments the “Arnoldian
concordat” that cut off “criticism" ftom “the creative” (CW 6). And
so on.]
Aims at least is quite
explicit about the ethical grounding of the enterprise. “The choice of an
interpretive norm is not required by the “nature of the text,” but, being a
choice, belongs to the domain of ethics rather than the domain of ontology”
(AIM 7). If we cannot depend on “neutral analysis in order to make decisions
about the goals of interpretation,” “we have to enter the realm” of
“ethical persuasion” (AIM 85), which presumably brings us back from logic to
rhetoric, the domain we saw disbarred before. "Even understanding"
itself “as contrasted with misunderstanding has only an ethical and not an
ontological claim to privilege" (AIM 135) -- again, a strongly performative
view of knowledge.
Hirsch says "this observation
had been made in the earlier book, but so briefly that it was generally
overlooked” (AIM 7), namely in the passage about “the interpreter”
“deciding what he wants to actualize” (VAL 25). But Validity is not
argued in ethical terms, except in its waspish asides, as when a scholar with a
“tolerance to a wide variety of readings" is denounced for “abject
intellectual surrender" and “abandonment of responsibility" (VAL
168). Still, the earlier book implied a systematic leap of faith from the
logical gradation of more versus less probable over to the ethical dichotomy of
right versus wrong. Few logicians would be inclined to designate classification
and hypothesis-testing as ethical operations. “Validation” procedures can at
best determine probability. The assertion of “correctness" is a crucial
additional performance on the critic's part, and one that steadily fewer critics
on the contemporary scene are eager to enact -- a stance which Hirsch may call
“unethical,” but hardly illogical. He raises his vehemence to arouse them
from a state which he considers lethargy and irresponsibility, but which may
well be a concern for other uses of literature than his.
As would be predicted, Hirsch
draws his most powerful ethical mandate from fidelity to the author. He now
calls the author's intended meaning the “original meaning,” and designates
it “the “best meaning," whereas any other is “anachronistic
meaning” (AIM 92, 77, 79, 88f). 16 [16. Hirsch says the
latter term is “a shorthand, not a pejorative term"; but it is the
converse of "the best'" (AIM 163, 92). This return to origins may be
yet another point of friction with the deconstructionists (cf. Notes 8 and 13)].
So the "ethical maxim for interpretation" is: “unless there is a
powerful overriding value in disregarding an author's intention (i.e., original
meaning), we who interpret as a vocation should not disregard it" (AIM 90).
Such an ethic presupposes that we know what the author's intention was; and the
ethical emphasis invites scholars to exaggerate the degree of their own
certainty on this point. An authorial orientation too readily becomes an authoritarian
one when the most probable reading becomes the “valid” one for which we must
surrender all others.
Hirsch is so convinced by his
own arguments that he can only comprehend the opposition of other critics as a
sign of “decadence” (AIM 13, 149). He accuses the “cognitive atheists,”
deconstructionists in particular, of two kinds of errors. In the first place,
“it is ethically inconsistent to batten on institutions whose very foundations
one attacks"; “it is logically inconsistent to write scholarly books
which argue that there is no point in writing scholarly books" (AIM 13).
But the implication that academics must believe what their institutional
overseers demand seems more apt for theologians (or commissars) than
disinterested scholars. And the “scholarly books," of which the
“atheists" produce more numerous and profound ones than Hirsch does, do
not, as we will see in Chs. 14-16, deny their own purpose, but their traditional
functions.
In the second place, he bars
“universal relativism” from claiming “absoluteness" in “a world
devoid of absolutes." But this argument falls apart too. Applied to itself,
relativism generates not “absoluteness,” but an ever-finer scaling of
greater or lesser relativity. By the same token, an increase in inconsistency
will not lead ultimately to consistency, but to greater diversity and nuancing.
Whatever logic may dictate, every system is operationally relative and
inconsistent in some aspects. At the conclusion of this chapter, I review the
inconsistencies in Hirsch's own position, wherein they are at least as
disruptive as any a deconstructionist would commit.
A better way to estimate Hirsch's claims than philosophical or ethical
disputation might be practical demonstration. Hirsch admits he has never
provided one for an entire literary work (interview with me, 1984). He is
content with sporadic illustrations whose simplicity conceals the problems I
have raised. His example for narrowing down classes is a dispute over two
competing versions of an Old English manuscript (VAL 187f). The objects of
contention (“ thwyrlic” versus “thrymlic”) were already produced by two
different scribes of the text and thus came ready-made for the interpreter to
adjudicate.
When his objects are
interpretive hypotheses, they also fall into neat, straightforward contrasts: an
optimistic versus a pessimistic way to construe Wordsworth's “A Spirit Did My
Slumber Seal'; death versus physical departure as the topic of Donne's
“Valediction Forbidding Mourning'; an Oedipal versus a non-Oedipal reading of
Hamlet; and so on (VAL 19Off, 122ff). Hirsch cheerfully concedes that his
demonstrations “work well” when one reading is “a sitting duck"
rather than an “expert reading” (VAL 192). But he sees only a quantitative
problem here that would be solved with more evidence – “all the known
relevant data." He does not foresee a qualitative jump in complexity and
indeterminacy that could undercut his methods.
In my estimation, the really
straightforward disputes tend to be rather uninteresting ones, hardly worth
amassing heaps of data. Consider Spenser's lines in Faerie Queene, (1,
ii, XiX); “And at his haughtie helmet making mark, so hugely stroke, that it
the steele did rive, and cleft his head.” I am uncertain whether the “it”
refers to “stroke” and is thus the grammatical subject, while the
“steele" is the foe's helmet; or to “helmet" and is thus the
direct object, while “steele” is the sword. 17 [17.This
example was brought to my attention by George Dillon's (1978) highly useful book
on linguistics and the reading of literature.] Gathering “all relevant
evidence” -- on 16th-century syntax, Spenser's usage, other passages in the
same work, and the like -- seems unreasonably arduous in light of what I stand
to gain by clarifying such a minor point. The “head” is thoroughly
“cleft” either way, and probing the syntax of the passage is, erm,
hair-splitting of a far less valorous kind.
Moreover, even straightforward
disputes may have to be settled in devious ways. Frost's poem about “The Road
Not Taken" says that one road “was grassy and wanted wear,” but just
afterwards that the “passing” “had worn them really about the same.” A
logician would object that both statements cannot be literally true at the same
time. To assume the author was just being careless would violate Hirsch's very
centre of authority (as we will see in a moment, he bends meanings every which
way to save some jarring passages of Blake's). If we undertake a validation, the
most relevant evidence would be found in the theme and lesson of the poem, both
of which convey the idea that “taking the road less travelled by" has
made "all the difference.” My students, who pick this poem a lot, usually
solve the conflict by assuming that the speaker is in a divided state of mind.
This interpretation, though reasonable enough if we want to rescue the point of
the poem, is reached not by virtue of verbal meaning of the incongruous
statement, but in spite of it.
In another famous poem, William
Carlos Williams says that “so much depends upon a red wheelbarrow beside the
white chickens." An anecdote has been preserved that this poem was inspired
when the author, a doctor by vocation, was gazing out from the window of a
gravely ill patient's house at a certain wheelbarrow (cf. Rhodes, 1965). This
fact might qualify as authorial evidence to “validate" the reading of
“so much" as “the chance of recovery, survival,” and so on. Yet the
evidence places the poem in so specific a context as to severely impoverish its
potential meaning. The poet certainly isn't asking us to believe that a
wheelbarrow cures illnesses if administered by chickens in white medical
uniforms. I would construe the connection in terms of how mundane objects can
seem vital and significant. This effect no doubt increases when the prospect of
death is imminent, but I do not get this specification from the poem as it
stands. So I am inclined to treat the anecdote as a curious sidelight, but by no
means a motive for enforcing a specific meaning.
These examples are still deceptively
elementary. I can scarcely imagine how Hirsch's method would deal with really
complex passages, such as:
And
the hapless soldier's sigh
Runs
in blood down palace walls (Blake)
A
breeze like the turning of a page
Brings
back your face: the moment
Takes
such a big bite out of the haze
Of
pleasant intuition it comes after (John Ashbery)
And
once below a time I lordly had the trees and leaves
Trail
with daisies and barley
Down the rivers of windfall light (Dylan Thomas)
and so on.
To “validate” interpretations for these, the critic has no tidy rows of
comparable “objects” to work with. Considerable skill would be needed to
manufacture the objects by translating the hypotheses of various readings into
some format that would group them under roughly similar categories. This process
would be interesting to observe, but I doubt it would resemble Hirsch's
“objective” methods very closely. Moreover, all the evidence we could gather
about the authors might only diversify rather than unify our hypotheses.
I could multiply examples
indefinitely, but my point would always be the same. The most important factor
Hirsch loses by denying the “special ontological status” of literature is
the experienced pressure to shape and reshape its “verbal meanings. “
Indeed, certain types of literature are expressly constituted by deliberately
abdicating authorial control over meaning. “Concrete" or “visual"
poetry presents configurations of letters and non-words rather like inkblots
designed for free association. “Found” poems depend on the determination of
the presenter and the reader to disregard the original author's intention and
“will.” Hirsch again prescribes an unintended and inappropriate response
when he suggests that “definitive interpretation” of a found poem requires
“access to the texts from which they were excerpted” (VAL 97). He refers us
back to the non-literary domain because his model discounts the special focus
that converts the “found” text into a poem.
I find it revealing that when
Hirsch does embark on large-scale interpretation, he by no means stays within
the limits of his own theory in his study of Blake's poetry. 18
[18. Hirsch cannot have simply changed his mind between 1964 (Blake study) and
1967 (VAL), because his original essay on “verification” was already
published in 1960. He admitted apologetically that he failed give a “model of
adjudication” for choosing between “two disparate modes of interpreting
Blake's Songs of Innocence and of Experience” because he did not
consider “all the important relevant evidence” (VAL 182).]
There, Hirsch has occasion to ask whether the phrase “Gave thee
clothing of delight” should be read as “God clothed the lamb with
delight," making “delight" a substance; or as “God gave the lamb
delightful clothing,” making “delight” a descriptive modifier -- and he
accepts both readings (IE 178). This sort of solution, spurned in his theory as
an “abandonment of responsibility” (VAL 168), seeks the “rich variousness”
he rejected in his argument on “determinacy” (VAL 45).
Hirsch may transform the
“verbal meaning” or contravene it if it collides with his reverence for the
author. In the Songs of lnnocence, the “childish simplicity of language
utterly belies their adult profundity of insight” (IE 21). Blake “must have
made an implicit moral judgment on cruelty and injustice which was at odds with
the benign and accepting surface tone of his poems” (IE 18). Whatever
authorial evidence Hirsch may have for these assertions must relate not to
“verbal meaning,” but to “significance.”19 [19. “Irony” is a
case where “two different mental sets” are “simultaneously adopted"
(AIM 107; cf. AIM 23ff). But since the two pertain to “verbal meanings"
in very different ways, I have to wonder once more how Hirsch's validation might
proceed.]
As we see, Hirsch finds it
expedient to interpret by overstepping the restrictions prescribed in his theory
for validation. He expounds “significance," "implications,"
“unsaid meaning,” "symbolic meaning,” "sacramental meaning,”
“prophetic meaning,” “images,” and "imaginative identification”
(IE 42, 30, 30, 43, 38, 40, 22, 249). He reminds us that Blake's poems are not
uniform objects: “each poem for him was a new start” and thus must not be
made an “intellectual counter" in a “dialectical system” (IE 5ff) -- yet his own logic of classes is presumably one such
“system.” And he ventures interpretive statements that could not be
validated in any logic, because their own verbal meanings contain stark
contradictions, for example: "Man is a Child, and a Lamb, and to others, a
Shepherd. Ultimately, Shepherd and Sheep, Father and Child, are the same” (IE
29). The issue of synonymy, about which he was so confident, cannot be sensibly
raised here.
Also instructive is Hirsch's
fabrication of a “reader” who construes the text the way Hirsch does. We get
told what “the adult reader implicitly knows"; what “the sensitive
reader will feel;” or what “all sympathetic readers of the poem have
experienced" (IE 178, 246). Such attributions imply an immediately attained
consensus among readers, though his theory predicts just the opposite principle.
Or, Hirsch adopts the familiar expedient of standing aside to let the text act
out the meaning on its own, even though he does not subscribe to the
independence of the text asserted by the New Critics (VAL 11f) .20 [20.
The “intentional fallacy” gets redeemed by claiming that its “careful
distinctions and qualifications” got lost in a “false and facile dogma"
(VAL 12); but compare de Man's meticulous argument (BI 24fl). Validity is
dedicated to Wimsatt, as is, ironically, its antithesis, Bloom's Anxiety
of Influence.] For Blake's “Tyger” (IE 244-252), he delivers a detailed
hypothetical account of the poem in action, for instance: “these staccato
beats of controlled fury are succeeded by a stanza of immense calm that
enormously widens the imaginative range of the poem” (IE 248). This practice
goes against his precept that “the “text” says nothing at all” by itself
(VAL 14). And when he gets to his value judgment, he gives none of the
“explicit criteria” he prescribed for “evaluation” (AIM 108), but merely
opines: “it is the most inclusive poem Blake ever wrote,” and “its
spiritual scope is immense" (IE 252).
The point here is not that
Hirsch's readings are vulnerable. That risk must be taken in stride for any
poetry as abstruse and uneven as Blake's. The point is that such challenging
texts can scarcely be interpreted in any productive of revealing way with a
weighty apparatus of validation confined to "verbal meanings."
Hirsch's interpretations of Blake can claim validity because they succeed in
“making functional many elements of the mute text” (cf. VAL 190).
In sum, Hirsch's enterprise is
not an intrinsic and logical one after all. Rather, it is an attempt to impose
an extrinsic ethic upon a diffuse and diversified activity by forcing the latter
into a reduced, normatized, and idealized mould. “Clues" about an
author's “meanings” need not “compel a right guess” any more than just
reading the text and trying to formulate “coherence” (cf. VAL 240, 170,
194). Only the critic's prior interpretation will decide what the clues are and
what they prove. As Hirsch says, “the interpreter" “decides what he
wants to actualise and what purpose his actualisation should achieve” (VAL
25). The circularity of understanding within context can only be traded for a
greater circularity of selecting and using external clues: greater because the
author evidently did not think the text needed the clues, but did compose a
context by including what seemed relevant and necessary.
The usefulness of validation
seems limited to the kind of obstacles that confront editors of definitive
editions or compilers of variorum (cf. VAL xi; AIM 89). The method has failed to
gain wide application not because today's literary critics are prey to
"decadence" and “atheism," but because it assigns to the
interpreter tasks that demand extreme efforts for meager rewards. Besides, even
as pure theory, it is rife with inconsistencies: studying effects without
causes; narrowing interpretation by broadening the evidence; arguing for the
self-identity of meaning but relying on outside clues; seeing reading as both
orderly and disorderly, but writing as fully determinate; seeking uniform traits
while denying that literature has any; distinguishing between interpretation and
criticism yet saying they are always intermingled; defending the author's
intended meaning but disregarding the author's intentions about determinacy; and
so on. When Hirsch puts "true" or “correct” in the place of
“most probable,” he commits the “logical error" of “erecting a
stable normative concept” "out of an unstable descriptive one” (VAL
13). Even a devoted apostle of “validation” would be torn by contrary
pressures.
Hirsch's theoretical arguments
are also beset by relativism, the crime of which he convicts the "atheists.
" He claims that science and humanities have the same “cognitive
element" (AIM 149), but instead of stating laws, regularities, or testable
predictions, he prefers to look for “maxims" and “rules of thumb,"
which he says “cannot be relied upon in any particular instance" (VAL
203; AIM 59) -- so that his claims are insulated against refutation by
counter-examples.
However, the impossibility of
proof is used as a shield against arguments contrary to his own. "That one
man's verbal meaning is always necessarily different from another's" is a
“hypothesis” which cannot be "falsified by empirical tests";
"the inaccessibility of verbal meaning is a doctrine" whose
“falsity" "neither experience nor argument can prove";
"that an author's verbal meaning is inaccessible” is “an empirical
generalization which neither theory nor experience can decisively confirm or
deny'; and so on (VAL 39, 33, 180. Such hedges are apparently thought sufficient
to dispose of these arguments while discouraging empirical research that might
endanger Hirsch's own position. 21 21. Hirsch told me in 1978
that he really isn't happy unless he's being attacked, which may have to do with
his championing of unpopular stances and his neglect of substantive evidence.]
But the superiority of literary scholarship or philosophical logic over
empirical tests has itself yet to be proven.
In fact, current research in
cognitive psychology reveals some noteworthy parallels with Hirsch's
deliberations. His concept of the “intrinsic genre” is quite close to what
is now usually called the “schema'. 22 [22. In 1978, I
introduced Hirsch to the renowned cognitive psychologist Walter Kintsch, whose
empirical research on reading was at the forefront. What effects may have ensued
I have not assayed to ascertain.] For a comparative review of models in text
research at the time, see Beaugrande (1980-81, 1982a, 1982b).] The term
“schema” appeared marginally in his earlier book and prominently in the
later, where Piaget is cited (VAL 109; AIM 3Iff). And Piaget's "schema”
is explicitly compared to Gombrich's “genre” (AIM 32; cf. VAL 104). Perhaps
a more empirical orientation might eventually be reached, though several
important points of divergence should be kept in mind. So far, psychological
probes of the “schema” have hardly dealt with literary or poetic
communication. Authorship and writing have been probed in far less detail than
reading. And schema theory (including Gombrich) is less concerned with
validation than with broader issues of expectation and coherence.
We need to consider here the
interests of both the reading and the criticism of literature, whose fate is at
stake in all literary theory (Ch. 3). If established, Hirsch's methods could
severely worsen the already alienating gaps between the subject and object, and
between the trained critic and the ordinary reader. Hirsch never spells out what
we are to do with a “verbal meaning" once it has been validated,"
though he all too readily assumes that a “student's reading” is “probably
wrong” in comparison to “his instructors"“ (VAL 737f). Presumably,
what is validated gets duly published and enforced in critical discussions,
including those of classrooms and textbooks. I can think of no more baleful way
to close and devitalise the literary experience, and to project an image of
literature as reserved for authorized insiders. The danger is not merely that
Hirsch wants to “throw out the experienced work of art and retain only the
scholarly apparatus,” as Louise Rosenblatt (1978:110) has justly remarked; but
that the apparatus would marginally useful and eventually suffocating. It would
mandate what Wellek and Warren call “a fixity and rigidity alien to the nature
of poetic statement" (TL 190).
At the same time, the
ontological status of literature would be levelled by the doctrine that
"genuine knowledge in any field, that is, sharable and usable knowledge,
depends on the communication of propositions about reality” (AIM 77). The
study of literature would be hard to justify on that basis; we would seem to be
dealing with a medium so inefficient, if not misleading, as to scarcely reward
the interpreter's efforts. Already embattled on many fronts, criticism can
scarcely afford a stupendous investment in a mode of research whose specific
relevance for literature has yet to be convincingly demonstrated.
In the meantime, we have at least the performance of a theoretician attempting to found the purpose and prestige of literary studies on objectivity and ethical commitment to a certain form of truth. So far, he has not followed up with the far more strenuous performance of putting the program into action on a complex and large-scale work of literature. This lack provides a counterpoint to those theoretical critics, such as Fiedler, Jauss, Holland, Paris, Bloom, de Man, and Millett, who work their theories out in close correlation with the issues of confronting specific texts. Hence, validation remains only a goal, or rather an attitude about a goal we cannot finally judge. The trends of the times are adverse to the attitude for motives I have tried to outline. In a different scholarly climate, however, we might profit from observing the experiment of stringently attempting to put such procedures into practice for specific problems of validation, such as conflicting editions of a Shakespeare play.
Click here to
go to Critical Discourse Main Page