Francesca Ervas 7
Journal of Language & Translation 9-2
September 2008, 7-29
Davidson’s Notions of Translation
Equivalence
Francesca Ervas
Università Roma Tre
Abstract
The paper analyses the relationship of semantic equivalence as
described by Donald Davidson in his theory of meaning, showing its
limits above all in respect to language use in the contextual situation.
The notion of equivalence used by the “first” Davidson does not
successfully explain why some biconditionals are simply true and
why others, besides being true, offer the real translation of the
source sentence. The paper argues that the main limits of the
Davidsonian proposal, which lie in the very attempt to apply
Tarskian theory of truth to natural languages, are partially overcome
later by Davidson himself. Above all in his paper A Nice
Derangement of Epitaphs (1986), Davidson rejects the very idea of
an “invariance of meaning” and proposes a “second” notion of
equivalence, described as the research of momentary and always
changing points of convergence of interpreter and speaker,
depending on contextual information. This convergence is possible
because of a “deeper equivalence,” a common cognitive apparatus
that allows communication to take place. At any rate, as the paper
aims to demonstrate, this solution seems to simply shift the problem
on to another level of explanation. Once this level of “deeper
equivalence” is reached, there is too no explanation of exactly how a
translator can understand contextual implications in order to grasp
functional equivalence.
8
Davidson’s Notions of Translation Equivalence
Keywords: translation equivalence, radical interpretation,
translatability, meaning invariance
1. Semantic Equivalence in Davidson’s Theory of
Meaning
Davidson’s early papers, collected in Inquiries into Truth and
Interpretation (1984), attracted the attention of Translation Studies
after a heated and often misleading debate on linguistic relativism
and Quine’s thesis on indeterminacy of translation, which denied the
existence of any “plausible sort of equivalence relation however
loose” among sentences (27). Davidson’s famous argument on the
very idea of a conceptual scheme showed the impossibility of a
radical difference between speakers and a massive failure in
translation, in accordance with the common sense of the practice of
translation.1 According to the principle of translatability, there is a
“deeper equivalence”, a wide common background that makes
comprehension possible. This common base of communication
relies on the fact that speaker and interpreter share the same
cognitive apparatus. As Malmkjær noted, translation scholars
discussed the Quinean thesis on the indeterminacy of translation in
depth, but the later work by Davidson is largely unknown. His last
papers, especially A Nice Derangement of Epitaphs (1986), as this
paper will show, are remarkably interesting, because they propose a
linguistic theory that can be used as a valid conceptual tool for the
analysis of the translation process:
It is therefore good to see that references and reactions to
Quine's indeterminacy thesis have begun to re-enter translation
literature Unfortunately, references to Davidson's more
1
For a brief but insightful review of the role of analytical philosophy in translation
theory, see K. Malmkjær (1998: 8-13).
Francesca Ervas 9
optimistic view are rarer, and I do not believe that his later
writings – absent even from Benjamin – have yet been
absorbed by the community of translation scholars. This is a
pity since the later work stresses difference and the fluidity of
language to a degree which should make it impossible any
longer to misread Davidson as seeking to establish “an original
and archaic site of meaning” and “an unmediated access to the
world” (Malmkjær 1993: 135).
At the same time, translation scholars were searching for a
highly formalised model of translation that could provide a formal
explanation of translation phenomena and could offer real laws of
translation as tools for translation work. In order to do this, as
Catford claimed, “it is clearly necessary for translation theory to
draw upon a theory of meaning” (35). And Davidson’s theory of
radical interpretation seemed to be able to answer the philosophical
question on meaning, revealing the translation of any given sentence,
whether expressed in another language or in the same language of
the original sentence. The theory Davidson proposed, tried to get to
the root of the problem of meaning, providing a radical interpreter
with a corpus of information which is both necessary and sufficient
to understand his own interlocutor. It must give an explanation of a
potentially infinite number of sentences, starting from a finite basic
vocabulary and finite set of rules to be used by an interpreter who
has finite powers.2 Davidson claimed that if a theory wants to have
these features and be able to give for any given sentence s its
equivalent sentence p in any other natural language, then it needs to
2
“A satisfactory theory of meaning must give an account of how the meanings of
sentences depend upon the meanings of words. Unless such an account could be
supplied for a particular language, there would be no explaining the fact that, on
mastering a finite vocabulary and a finitely stated set of rules, we are prepared to
produce and to understand any of a potential infinitude of sentences” (Davidson
1984: 17).
10
Davidson’s Notions of Translation Equivalence
adopt Tarski’s theory of truth as formal model. Furthermore, the
theory needed to be empirically verifiable and capable of giving a
holistic, global explanation of how a natural language works.
The model of translation proposed by Davidson soon showed its
limits with regards to the explanation of translation phenomena,
precisely because of formal problems within the theory itself.
Despite its interesting attempt to borrow the concept of equivalence
from Tarski’s theory of truth to define translation in natural
languages, another problem with Davidson’s theory was the real
application of this concept of equivalence to translation in natural
languages. In fact, the concept of equivalence borrowed from
Tarski’s theory of truth soon seemed too rigid to explain the various
kinds of equivalence that Translation Studies pointed out in relation
to the various features of natural language that remain constant in
real cases of translation.3 For example:
(1) Chi non risica, non rosica. (Italian)
Nothing ventured, nothing gained.
(2) Ho preso due piccioni con una fava. (Italian)
I killed two birds with one stone.
Referring to the conditions of truth of these sentences does not
explain why, in the first example, we can consider the second
sentence a translation of the first, even though the translation does
not respect an acoustic equivalence. Nor can it explain why in the
second example we can consider the second sentence a translation of
the first one, just because it respects a functional equivalence. In this
theoretic context, we lost the possibility of explaining a great variety
of translation phenomena linked to other kinds of equivalence of
3
For an introduction to the concept of translation equivalence, see Koller (1989:
99-104), Draskan (1986), Jäger & Neubert (1982), Newman (1994: 4694-4700),
Snell-Hornby (1995: 13-22), Halverson (1977), and Kenny (1998).
Francesca Ervas 11
some properties of natural languages that Davidson’s theory cannot
explain.
One of the greatest difficulties, as Tarski himself highlighted, is
the fact that natural languages are too unstable, confusing and
complex to apply formal methods to them. In natural languages, a
single expression can have more than one meaning, a not-welldefined meaning, or it can contain some indexical terms that make it
true or false depending on the context. So it would not be possible
even to presuppose a complete dominion of the syntax of an objectlanguage, which, as in any natural language, does not have clearly
defined outlines. Furthermore, it is impossible to establish which
new words might be added to the finite vocabulary of a natural
language: natural languages are not “static” like formalised
languages, instead they are in constant evolution, because new
expressions or words can always be added. So the worry of
translation scholars was, as Tarski had foreseen, that “the language
of everyday life, after being ‘rationalized’ in this way, would still
preserve its naturalness and whether it would not rather take on the
characteristic features of the formalized languages” (1956: 267). In
order to apply an appropriate Tarskian analysis of natural languages,
Davidson should have demonstrated first that Tarskian methods can
actually be applied to natural languages.
Despite the fact that Davidson’s theory did not include the
explanation of many cases of translation equivalence in real
translations, we could search for a theory that explains why a
translation is a translation in general, without referring to the issue
of a translation’s degree of quality. In the previous cases, we can
state that the first translation is a bad translation from the point of
view of acoustic equivalence, while the second case is a good
translation from the point of view of functional equivalence. But
both of them are translations. Translation scholars needed to
understand why such translations can be considered translations at
all, irrespective of whether they are good or bad. Above all, they
12
Davidson’s Notions of Translation Equivalence
wanted to understand why something can be considered a translation
despite its apparent lack of sameness of meaning, as in the second
case we considered.
Moreover, the Tarskian model helped Davidson give an
explanation of the meaning and thus of the sameness of meaning in
terms of truth conditions, avoiding intensional notions like those of
meaning and synonymy (having the same meaning), which did not
seem to “oil the wheels of a theory of meaning” (Davidson 1999:
20-21). In fact, the theory Davidson aimed to build can neither
presuppose nor refer to meanings of sentences, because it has to
provide an exact explanation of them, so what is required for a
theory of meaning is a characterisation of the meaning of sentences
by a strictly extensional formulation. At any rate, as this paper
intends to demonstrate, Davidsonian theory of meaning failed to
give an explanation of what translation is, even with regard to very
simple sentences from a syntactic and a semantic point of view.
According to his extensional approach to meaning, Davidson
believed that the meaning of a sentence can be explained “in terms
of truth,” indicating its truth conditions (Hacking 1975: 173).
Because we cannot assume the sentence on the right side of a Tsentence has the same meaning (nor that it is a translation) of the
sentence on the left side, the meaning of the sentence on the left side
might be understood by indicating the truth conditions in the
sentence on the other side. The radical interpreter can indicate the
truth conditions in order to understand the meaning of foreign
utterances, because he already has a certain understanding of the
“ordinary” concept of truth. So he already knows when the Tsentences are true. But it was pointed out that in this way there could
be some biconditionals the translator recognises as true and correct
from an extensional point of view, even though they do not offer the
translation of the sentence he wants to translate.4 To say that two
4
See criticisms in Hacking (1975: 142-145), Segal (1999: 48-56), and Foster
Francesca Ervas 13
sentences that make up a true T-biconditional are equivalent means
that they have the same truth-value. Two sentences which imply
each other are equivalent. Sentences with the same meaning are
equivalent, and yet equivalent sentences can differ in meaning!
Therefore the translator would know that all the following
biconditionals are true:
(3) “La neve è bianca” is true if and only if snow is white.
(4) “La neve è bianca” is true if and only if grass is green.
(5) “La neve è bianca” is true if and only if snow is white and 2
+ 2 = 4.
But the translator would know only that they are true
biconditionals and, without knowing the meaning of the foreign
sentences, he would not be able to say which biconditional
effectively is the translation of the sentence “La neve è bianca”. He
would not know how to translate it at all! The solution is not found
by claiming that the correct biconditional on the right side provides
the translation of the sentence on the left side, because we cannot
presuppose the concept of translation.5 So the corpus of information,
given by Davidson to the translator, would not be sufficient to
guarantee the choice of the theory that implies the semantically valid
T-sentence.
According to Davidson, we can recognise at least some analogies
between the structure of the sentence “La neve è bianca” and the
sentences in English “Snow is white” or “Grass is green”.
Eventually we would prefer the biconditional that associates “La
neve è bianca” to the sentence “Snow is white”, which “we have
5
(1976: 1-32).
“What we require, therefore, is a way of judging the acceptability of T-sentences
that makes no use of the concepts of translation, meaning, or synonymy, but is
such that acceptable T-sentences will in fact yield interpretations” (Davidson,
1984: 150).
14
Davidson’s Notions of Translation Equivalence
good reason to believe equivalent” (1999: 26). If by the term
“equivalence” we mean “semantic equivalence,” it would be
difficult for the translator as a radical interpreter to recognise this
“equivalence” between the sentence on the right side and the one on
the left side of the biconditional. In fact, from the beginning,
Davidson presupposed that the translator does not know the
meaning of the sentences, so he cannot say whether there is such an
equivalence. But in a situation of radical interpretation, the
interpreter cannot suppose even a syntactic-grammatical equivalence
between the sentences of the foreign speaker and the sentences of
his own language. In fact, the radical interpreter cannot draw any
analogy in a situation like the following:
(6) “Skuppit gromper” is true if and only if rubies contain coal.
In this case, as pointed out by J. Heal, the radical interpreter
could believe “Skuppit gromper” is a disordered set of sounds
without meaning (1997: 175-195). He has no possibility of
understanding in which sense the two sentences can be equivalent.
2. In Search of the Semantically Correct
Equivalent
Davidson proposed to solve this specific problem by finding the
biconditional that offers a real equivalent sentence in another
language, the semantically correct T-sentences for any sentence of a
language, through an empirical verification based on “facts about
the behaviour and attitudes of speakers in relation to sentences”
(1984: 133) He identified two different moments: a verification
before the interpretation has begun and a successive verification led
by the Principle of Charity. This passage was particularly interesting
for translation scholars because it shows how the theory can be
Francesca Ervas 15
verified in real cases of translation and how they can be generalised
in order to obtain laws of translation which allows us to predict
future cases. In any case, this attempt encounters another series of
difficulties.
According to Davidson, empirical evidence available to a radical
interpreter before the interpretation has begun is the attitude of
holding true a sentence by a speaker x at a given time t. But is it
possible to identify this attitude without referring to any theory of
interpretation? It is not easy to establish when the interpretation
begins if, as Davidson claims, the interpretation of sentences is
strictly bound to the interpretation of beliefs and intentions of a
speaker; or if it is part of a single interpretative project which
unifies linguistic behaviour and intentional actions of a speaker
through holistic criteria. If we adopt this point of view, the
acknowledgement of a speaker’s attitude of holding true a sentence
already requires an interpretative act by the interpreter. If everything
is holistically included in a single project, we cannot say that the
interpreter can recognise a speaker’s intentional attitude before the
interpretation has begun. How is it possible then to say that a radical
interpreter holds true a sentence without “having any idea of what
truth” (Heal 1997: 135).6 is and without knowing what the sentence
means and what the speaker’s beliefs are?
To identify a single attitude of holding true a sentence, the
radical interpreter must be able to distinguish voluntary from
involuntary behaviour and linguistic from non-linguistic behaviour.
He can only make these distinctions if he has some hypotheses that
involve the speaker’s beliefs and intentions. Furthermore, holding
true a sentence is something that can be done within the background
of a set of a community’s linguistic (and non-linguistic) practices
the radical interpreter does not know.7 So if we want to identify the
6
7
On the attitude of holding true a sentence, see Davidson (2001: 137-153).
See criticism in Picardi (1989).
16
Davidson’s Notions of Translation Equivalence
attitude of holding true a sentence, we need to imagine a situation
which differs greatly from that of Davidsonian situation of radical
interpretation. The hypothesis of basing interpretation on the
speaker’s attitude of holding true a sentence is too weak for the
interpreter to understand the speaker without knowing his language
or beliefs.
Nevertheless, even if we assume a situation which is different
from that of a radical interpreter, that is to say a situation which is
more similar to a real translator, what remains to clarify is how to
use this set of sentences held true by different speakers in different
situations as evidence to support the semantically correct Tsentence, the real translation. If the successive verification is based
on publicly available evidence, it is not possible to give an overall
explanation of how natural language works, because we cannot
consider many cases of linguistic utterances for which there is no
publicly available data (e.g., “Gods live in Olympus”). At any rate,
it is clear that even if the empirical check succeeds in a significant
sample of cases, we can never be sure that the sentence placed on
the right side of the T-sentence gives the meaning of the sentence
placed on the left side. In fact, if a sample of n confirmed cases
gives us the verification we are looking for, how can we be sure the
next case (n+1) will also be confirmed? The constraints imposed on
Davidson’s theory do not guarantee the passage from the
verification of n cases to the verification of the next case n+1. But
the translator must be able to predict future cases to understand the
meaning of a potentially infinite number of sentences the speaker
may utter.
In Foster’s opinion (1976), in order to guarantee that the theory
offers us T-sentences which actually give the meaning of the
speakers’ expressions, it has to provide the interpreter with some
prescience of the meaning of the sentences. But Davidson’s theory
comes with some biconditionals containing mere regularities which
cannot be projected on to still unobserved cases. The biconditionals
Francesca Ervas 17
do not establish a link of nomic character between the “left”
sentence and the “right” one. From Foster’s point of view, if such
biconditionals are not laws, they cannot support the appropriate
counterfactuals. In the essay Reply to Foster (1976), Davidson
accepts Foster’s remarks to solve the problem of extensionality.
Both the biconditional (7): “‘La neve è bianca’ is true if and only if
snow is white” and the biconditional (8): “‘La neve è bianca’ is true
if and only if grass is green” are recognised as correct by Davidson’s
extensional theory, even though only the first is semantically valid.
In Davidson’s opinion, we could solve this problem by recognising
the nomological character of T-sentences. In fact, the biconditional
(8) is not a law, because it cannot support the appropriate
counterfactuals. The counterfactuals in question would respectively
be:
(7) “La neve è bianca” wouldn’t be true if snow weren’t white.
(8) “La neve è bianca” wouldn’t be true if grass weren’t green.
In the second counterfactual, there could be a possible different
world, where grass is not green but snow is white. But the
nomological solution also is not free from difficulties. For instance,
there could be some worlds where “Snow is white” does not mean
that snow is white, but that grass is green, because there are different
linguistic conventions from ours. In such a case, the biconditional
(8) would be correct.8 Even if we concede to Davidson that the
biconditional (8) cannot be considered a law in his system, the
problem of extensionality is not completely solved. In fact, by
means of the nomological solution he did not succeed in eliminating
another kind of biconditionals seen above:
(9) “La neve è bianca” is true if and only if snow is white and 2
+ 2 = 4.
8
Compare the Introduction in Fodor & Lepore (1992: III).
18
Davidson’s Notions of Translation Equivalence
Davidson’s response to this objection was that his theory of
interpretation, “being an empirical theory, favours simplicity, like
any other empirical theory” (1999: 57). Therefore, according to
Davidson, this last biconditional should be rejected, because by
adding irrelevant clauses, the criterion of simplicity imposed on
theory would not be respected. But who is to judge the simplicity of
the biconditionals, and on what basis? The concept of simplicity is
too relative, because it would involve the consideration of many
different opinions and interests. The requirement of simplicity does
not solve the problem; what “simplicity” might mean is too vague.
Such a requirement would not be respected in any case, because we
could build a T-sentence and thus a law for every sentence of an
object-language. But as we know, Davidson proposed that sentences,
though constructed with a finite vocabulary and system of rules, are
potentially infinite. How could a theory that builds a law for every
linguistic utterance of a speaker be “simple”? Davidson responds
that the kind of theory he has in mind does not have to imply
nomological T-sentences for every sentence of a language. But then
he does not clarify which sentences of language have to possess
their correspondent T-sentences being of nomological character and
which, on the contrary, have to remain simple T-sentences. In this
way, any mere accidental regularity and even any single instance
could be a law! For example, we could build these two laws:
1. “All the bars of gold are less than one km long” is true if and
only if all the bars of gold are less than one kilometer long.
2. “This bar of gold is less than a km long” is true if and only if
this bar of gold is less than a kilometer long.
As we can see, any instance of what Davidson considers a law
will in turn be a law! And we could build a law for a real physical
law in the same way:
Francesca Ervas 19
3. “All the uranium bars are less than a km long” is true if and
only if all the uranium bars are less than a kilometer long.
As we know, a one-kilometer-long uranium bar cannot exist,
because it would own a mass superior to the critical mass. According
to physical laws, if uranium 235 overcomes its critical mass, a
nuclear reaction follows. These three examples highlight different
situations. Yet in Davidson’s opinion, all three biconditionals would
be laws without distinctions. In this way, Davidson’s nomological
solution places single instances and laws on equal level!
3. Prior and Passing Theories
Davidson, in the last part of Inquiries into Truth and
Interpretation (1984), Limits of the Literal, takes the limits of the
application of his theory of meaning into consideration and
recognises that “it is always an open question how well the theory
an interpreter brings to a linguistic encounter will cope” (xix).
According to Davidson, the field of application of his theory of
meaning should be very wide to give a comprehensive explanation
of how a language works. Yet at the same time it should be narrow
enough to be rigorously formalised. However, missing in such a
formalisation is exactly what helps us to apply the theory to a single
case of interpretation, because it cannot be reduced to a clearly
defined set of rules. In other words, an interpreter maintains the
conversation by continually adjusting his theory of interpretation by
means of his ability and intuition, but he is also aided by factors,
such as luck, taste and sensitivity. But he cannot expect to succeed in
formalising such considerations which lead to the adjustment of his
theory in accordance with the latest incoming information.9
9
Cf. Davidson (1984: 279).
20
Davidson’s Notions of Translation Equivalence
Interpretation is a gradual process led by the interpreter’s ability
to adjust his own solutions to what the speaker seems to believe and
mean. This is the main thesis claimed in Communication and
Convention (1981) and in A Nice Derangement of Epitaphs (1986).
Here Davidson attempts to explain “how people, who already have a
language manage to apply their skill or knowledge to actual cases of
interpretation” (1986: 441). The search for equivalence is described
as an attempt to continually bring into action our expectations in the
single communicative encounter in accordance with incoming
information, to obtain an actual and concrete agreement with our
interlocutor in translation. This interpretive process is possible
because of a common background, a “basic agreement” or a “deeper
equivalence” expressed by Davidson’s principle of translatability.
According to this principle, translatability is guaranteed by a “basic
cognitive apparatus” which is common to all human beings and
enables the interpreter to detect the similarities in the speaker’s
utterances which make it possible for them to understand each other.
Interpretation is no longer a mechanical process governed by a
clearly defined set of rules and conventions learned before their
infinite cases of application. Davidson himself compares such
theory to a “machine” that seems to make language a
complex abstract object, defined by giving a finite list of
expressions (words), rules for constructing meaningful
concatenations of expressions (sentences), and a semantic
interpretation of the meaningful expressions based on
semantic features of individual words. We tend to forget
that there are no such things in the world; there are only
people and their various written and acoustical products
(2001: 107-108).
Now Davidson concedes that we never need this kind of
language in our everyday communication with others, even though it
Francesca Ervas 21
could still be interesting to philosophers, psychologists and linguists.
We simply understand what other people tell us, and we manage to
be understood without such an unobservable and unchangeable
“object”. Davidson concludes “there is no such thing as what some
philosophers (me included) have called a language” (1994a: 2). He
questions the image of language he shared for a long time with most
linguists and philosophers of language. Now he is more interested in
phenomena such as lapsus linguae and malapropisms, which
standard descriptions of linguistic competence do not take into
consideration, because – according to Davidson – they are unable to
explain how our communication succeeds, despite the presence of
such phenomena.
Davidson should have abandoned the concept of equivalence he
borrowed from Tarski’s theory of formalised languages, because
this concept, while suitable for a complex and abstract language
built on the model of languages of deductive sciences, is not suitable
for spoken languages in everyday life. Nevertheless, Davidson does
not want to refute his former theoretical programme with this more
recent paper and its theses.10 On the contrary, he intends to claim
that a theory of meaning, such as the one he elaborated previously,
can be used to describe the linguistic behaviour of speakers in a
systematic and coherent way, rather than to explain the link between
such a theoretical “machine” and our practical interest in
understanding and being understood in every single case of
interpretation, even where the phenomena stated by the theory are
divergent.
Davidson claims that the interpreter interprets a speaker with an
interpretative theory (again based on Tarski’s model) called prior
theory, which illustrates how the interpreter is already prepared to
interpret the speaker (the first meaning). The prior theory always
refers to a given speaker placed in a given situation, a speaker about
10
For criticism of Davidson, see Hacking (1986).
22
Davidson’s Notions of Translation Equivalence
whom the interpreter knows little (for example, his dress, sex, etc.) –
information acquired by an initial observation of his behaviour – or
has deeper knowledge. Therefore there will be an infinite number of
prior theories in accordance with the speakers, the interpreter's level
of knowledge about the speaker’s customs, socio-cultural situation
and background. During this interpretative process, the interpreter
will modify his initial theory in accordance with the entry
information, building one or more passing theories which express
how the interpreter actually interprets the speaker. Davidson quotes
as an example the case of Mrs. Malaprop, who, by saying “a nice
derangement of epitaphs,” actually meant to say “a nice arrangement
of epithets”: her interpreter has to use a new theory assigning a new
meaning to “derangement” and “epitaph”.
Yet simply because the interpreter applies a passing theory to
understand the speaker, he cannot be certain that his accommodation
of the prior theory will work for every other future utterance of that
speaker, because, according to Davidson, a passing theory tells us
only how we have to “interpret a particular utterance on a particular
occasion” (1986: 443). So the prior theory will be continually
modified with other passing theories, it will be improved through
constant comparison with the speaker.11 Ultimately, the interpreter
will understand the speaker through this constant, progressive
adjustment of his own theories. In this way, Davidson underscores
the interpreter’s creativity, which cannot be explained by a
description of linguistic competence as a finite set of rules,
conventions or uses determined by history, linguistic practices, etc.,
to apply to specific cases. 12 Davidson intends to abandon such
explanation of linguistic competence to stress instead the central
role of creativity in language use. In fact, he maintains that constant
accommodation of a prior theory and the use of passing theories to
11
12
Cf. Davidson (1986: 437).
Cf. Davidson (1986: 443).
Francesca Ervas 23
understand the speaker’s utterances characterise not only those
situations where we have to understand lapsus linguae or
malapropisms, but every interpretation, because he thinks phenomena
such as lapsus linguae or malapropisms are “omnipresent and
pervasive” (1986: 433). Therefore every communication needs an
accommodation of the prior theory and a constant change of passing
theories. Understanding can also take place without rules, uses or
shared practices; maybe it is facilitated by these conventions, but
what makes it possible is the creative dimension of our linguistic use.
But linguistic phenomena such as lapsus linguae or
malapropisms do not seem omnipresent; they are usually considered
exceptional cases in our usual way of communicating. Speakers do
not usually mistake their own words, they simply say what they
mean. So we cannot totally abandon the descriptions of language
that rely on rules and conventions, because these phenomena are not
“rules” but “exceptions”. As Dummett claimed, conventions teach
us:
what constitutes a social practice; to repudiate the role of
convention is to deny that a language is in this sense a
practice. In the exceptional cases there are indeed no rules to
follow: that is what makes such cases exceptional (1986: 474).
On the contrary Davidson, emphasises the creative and
productive power of language, giving as an example, in the paper
James Joyce and Humpty Dumpty, Joyce’s use of language. Joyce
abandons his nationality, religion and language, not to annihilate his
language, but to recreate it, putting his reader “in the situation of the
jungle linguist trying to get the hang of a new language and a novel
culture” (1991: 11). Joyce’s “radical reader” somehow understands
the meaning of what Joyce writes, even though he strays outside the
rules and conventions of his own society and language. But we can
also claim that Joyce’s way of writing could be considered an
24
Davidson’s Notions of Translation Equivalence
“exceptional” case and not a typical case of communication. It is not
so easy to understand how, without any rules, conventions, or social
practices, the interpreter can understand the speaker in a real case of
communication.
However, Davidson does not believe that Joyce’s language is
absolutely without ties, as it was created ex nihilo. Davidson
disagrees with Humpty Dumpty: “When I use a word it means just
what I choose it to mean” (1991: 1), because he does not believe
in the existence of a private language. As Wittgenstein claimed,
without sharing one’s own language with anybody else, the speaker
cannot possibly know what the correct language use is. Instead the
language created by Joyce is intersubjective, and – however
detached from rules and conventions it may be – it opens “a
hermeneutic space between the reader and the text” (1991: 12),
which is shared by both speakers.
4. A “Deeper Equivalence”
In 1986, Davidson proposed a theory of meaning based on the
notion of linguistic use, to explain many translation cases (such as
those of “functional equivalence”) which remain unexplained by the
previous Davidsonian theory. In the words of Malmkjaer: “The
theory Davidson advocates provides a method and a concept of what
meaning is, which allows us to make sense of the linguistic and
other behaviour of other persons, and to see how their use of certain
sentences relates to their use of certain other sentences” (2005: 56).
The past usage of sentences could surely help the interpreter/
translator, because they become a background against which
linguistic items participate in semantic relationships formed by the
momentary fusion of speaker, listener and contextual situation.
Viewed from this perspective, we could describe translation as a
process that occurs between the original text and a target text
Francesca Ervas 25
through a series of convergences of passing theories. The bridge
between the two sides of the process is the translator himself, who is
the reader of the original text and the writer of the translation. In real
cases of communication, functional equivalence becomes a
momentary, always changing agreement between the interpreter or
the translator and the speaker. In this sense, the translator’s role
becomes that of “a mediator whose task it is to facilitate
convergence on passing theories for people who do not share what
we normally think of as ‘the same language’” (1993: 146).
But Davidson’s attempt to describe functional equivalence seems
to introduce a series of solutions ad hoc, rather than a real theory of
meaning. Moreover, this solution does not avoid the problems of
Davidson’s original theory, as examined above. As we have seen,
Davidson did not abandon Tarski’s theory of truth at all, instead he
tried to reconcile it with the new requirements for a theory of
meaning. Moreover, the momentary convergence is only an
equivalence of a long series of moments which constitutes the
translation process, and it is made possible by a “deeper” agreement
among speakers: the sharing of the same human mental dimension.
It is this “more fundamental equivalence – Benjamin claims – which
in turn engenders the possibility of the recognition of semantic
equivalence” (1989: 64-65).
Davidson refuses any form of conceptual relativism, rejecting the
distinction between conceptual scheme and empirical content. In the
essay On the Very Idea of a Conceptual Scheme (1974), Davidson
tries to demonstrate that relativism and in particular the thesis of the
radical difference between conceptual schemes, is bound to fail on
its own. The thesis of the incomparability of radically different
conceptual schemes is belied, in Davidson's opinion, by the
metaphor used by the advocates of conceptual relativism. According
to this metaphor, we can compare conceptual schemes to points of
view that give us a radically different vision of reality:
26
Davidson’s Notions of Translation Equivalence
The dominant metaphor of conceptual relativism, that of
differing points of view, seems to betray an underlying
paradox. Different points of view make sense, but only if
there is a common co-ordinate system on which to plot
them; yet the existence of a common system belies the claim
of dramatic incomparability (1984: 184).
Davidson believes that it is possible to talk sensibly about the
diversity of varying points of view (conceptual schemes) only if the
latter can be placed into a “system of common coordinates” in
which it is possible to compare them. Davidson concludes that, if
these points of view have something in common, they cannot be so
radically different! This means that the same recognition of
similarity and difference which occurs in translation is made
possible because of a fundamental cognitive identity among human
beings. Therefore, translatability is guaranteed by the fact that there
is human communication.
However, this solution seems to simply move the problem of
explaining functional equivalence to another level of explanation. If
this fundamental identity can explain the “deeper” equivalence by
means of which translation is always possible, it is not sufficient to
explain the evident differences in translation quality. In order to
explain why some translations seem to be more adequate than others
in a contextual situation, it is necessary to refer also to the
fundamental difference and individuality of the subjects who
translate, to the translator’s creativity underscored by the “last”
Davidson. In this way, however, what remains to be explained is
exactly how the translator can create something new by grasping the
functional equivalence of sentences of the source text and sentences
of the target text. Although Davidson recognises the importance of
the translator’s creativity, he remains anchored to the relation of
equivalence proposed in his theory of radical interpretation and is
unable to fully explain the creativity of the relation between the
Francesca Ervas 27
interpreter and the speaker. The fact of sharing a “basic cognitive
apparatus” can certainly explain the fundamental reason why
translation is always possible; however, it does not explain why
translations differ one from the other, and why we consider some
translations better and more creative than others.
References
Benjamin, A. 1989. Translation and the Nature of Philosophy: A
New Theory of Words. London & New York: Routledge.
Catford, J. 1965. A Linguistic Theory of Translation: An Essay on
Applied Linguistics. Oxford: Oxford University Press.
Davidson, D. 1984. Inquiries into Truth and Interpretation. Oxford:
Oxford University Press.
_____. 1986. A Nice Derangement of Epitaphs. In E. Lepore (ed.),
Truth and Interpretation. Perspectives on the Philosophy of
Donald Davidson 433-446. Oxford: Blackwell.
_____. 1991. James Joyce and Humpty Dumpty. In P. French, T.
Uehling & H. Wettstein (eds.), Midwest Studies in Philosophy
16: 1-12. Notre Dame: University of Notre Dame Press.
_____. 1994a. The Social Aspect of Language. In B. McGuinness &
G. Oliveri (eds.), The Philosophy of Michael Dummett 1-16.
Dordrecht: Kluwer Academic Publishers.
_____. 1994b. Dialectic and Dialogue. In G. Preyer, F. Siebelt & A.
Ulfig (eds.), Language, Mind and Epistemology: On Donald
Davidson's Philosophy 429-437. Dordrecht: Kluwer Academic
Publishers.
_____. 1999. Reply to Gabriel Segal. In U. Žeglen (ed.), Donald
Davidson. Truth, Meaning and Knowledge 57-58. New York:
Routledge.
_____. 2001. Subjective, Intersubjective, Objective. Oxford:
Clarendon Press.
28
Davidson’s Notions of Translation Equivalence
Draskau, J. 1986. The Quest for Equivalence. Copenhagen:
Athenium.
Dummett, M. 1986. A Nice Derangement of Epitaphs: Some
Comments on Davidson and Hacking. In E. Lepore (ed.), Truth
and Interpretation. Perspectives on the Philosophy of Donald
Davidson 459-476. Oxford: Blackwell.
Fodor, J. & E. Lepore. 1992. Donald Davidson: Meaning Holism
and Radical Interpretation. In Holism: A Shopper’s Guide.
Oxford: Blackwell.
Foster, J. 1976. Meaning and Truth Theories. In G. Evans & J.
McDowell (eds.), Truth and Meaning 1-32. Oxford: Oxford
University Press.
Hacking, I. 1975. Why Does Language Matter to Philosophy?
Cambridge: Cambridge University Press.
_____. 1986. The Parody of Conversation. In E. Lepore (ed.), Truth
and Interpretation. Perspectives on the Philosophy of Donald
Davidson 447-458. Oxford: Blackwell.
Halverson, S. 1997. The Concept of Equivalence in Translation
Studies: Much Ado about Something. Target 9. 2, 207-233.
Heal, J. 1997. Radical Interpretation. In C. Wright & B. Hale (eds.),
A Companion to the Philosophy of Language 175-195. Oxford:
Blackwell.
Hjort, A. 1990. Translation and the Consequences of Scepticism. In
S. Bassnett & A. Lefevere (eds.), Translation, History and
Cultures 38-46. London: Pinter.
Jäger, G. & A. Neubert. (eds.). 1982. Äquivalenz bei der Translation.
Leipzig: Enzyklopadie.
Kenny, D. 1998. Equivalence. In M. Baker (ed.), Routledge
Encyclopaedia of Translation Studies 77-80. London & New
York: Routledge.
Koller, W. 1989. Equivalence in Translation Theory. In A.
Chesterman (ed.), Readings in Translation Theory 99-104.
Helsinki: Finn Lectura.
Francesca Ervas 29
Malmkjær, K. 1993. Underpinning Translation Theory. Target 5.2,
133-148.
_____. 1998. Analytical Philosophy and Translation. In M. Baker
(ed.), Routledge Encyclopaedia of Translation Studies 8-13.
London & New York: Routledge.
_____. 2005. Linguistics and the Language of Translation.
Edinburgh: Edinburgh University Press,
Newman, A. 1994. Translation Equivalence: Nature. In R. Asher &
J. Simpson (eds.), The Encyclopedia of Language and
Linguistics 4694-4700. New York: The Oxford and Pergamon
Press.
Picardi, E. 1989. Davidson on Assertion, Convention and Belief. In
J. Brandl & W. Gombocz (eds.), The Mind of Donald Davidson
97-107. Amsterdam: Rodopi.
Quine, W. 1960. Word and Object. Cambridge, MA: MIT Press.
Segal, G. 1999. How a Truth Theory Can Do Duty as a Theory of
Meaning. In U. Žeglen (ed.), Donald Davidson. Truth, Meaning
and Knowledge 48-56. New York: Routledge.
Snell-Hornby, M. 1995. The Illusion of Equivalence. In Translation
Studies. An Integrated Approach 13-22. Amsterdam: John
Benjamins.
Tarski, A. 1956. The Concept of Truth in Formalized Languages. In
Logic, Semantics, Metamathematics 152-278. Oxford: Clarendon
Press.