Interpreting Saussure’s differential theory:

Merleau-Ponty versus the linguists*

Lei Zhu
Shanghai International Studies University**
zhulei@[Link]

ABSTRACT. Saussure’s differential theory, growing out of his historical

linguistic practice and ripening in his Course in General Linguistics, states
that no sound or concept exists prior to the linguistic system, in which
there are only phonetic and conceptual differences. This idea inspired
both Merleau-Ponty and the post-Saussurean linguists: the former found it
a useful tool in achieving his phenomenological goal without incurring a
subjective tendency; the latter made it a principle in the understanding of
phonemes and linguistic facts in general. However, Saussure did not
explain how the elements of a structure could actually arise from pure
differentiation, and this caused a problem for both Merleau-Ponty, who
felt it necessary to account for the process as a phenomenologist, and the
linguists, who were compelled to consider the process in their abstraction
of phonemes from raw data. To solve the problem, Merleau-Ponty turned
to his own distinction between the “speaking language” and the “spoken
language” as well as Saussure’s “speech”, and highly stressed the act of
speaking as the locus where concepts are generated through
differentiation. In contrast, the linguists tried to look beneath structural
differentiation for more “concrete” and “reliable” guarantees, which they
finally found in the atom-like phonetic properties or “distinctive features”,
understood as physical existences. Thus, structural differentiation in
linguistics was actually reduced to pre-existing “things” and lost its
originary status.

* This research was supported by Shanghai Peak Discipline Program (Class I): Foreign
Languages and Literature
** Correspondence: Lei Zhu – Institute of Linguistics (letterbox 338), Shanghai International
Studies University, 200083 Shanghai, 550 Dalian Rd. (W.), China.

Metodo Vol. 4, n. 2 (2016)

 90 Lei Zhu

1. Introduction
Merleau-Ponty’s encounter with structuralism in the mid-20th century
has been known to be one of double significance. It changed
phenomenology, in that structuralism became an essential source of
ideas in Merleau-Ponty’s phenomenological exploration outside the
existentialist paradigm represented by Sartre. It also changed
structuralism, to the extent that Merleau-Ponty, as «probably one of
the first French philosophers to become interested in Saussure», 1
played a central role in the early dissemination of structuralist ideas.
Despite these changes, however, it is equally known that the
extensive dialogue and rapprochement between phenomenology and
structuralism (the latter represented by the various human sciences of
the time) which Merleau-Ponty had looked forward to did not really
happen. Instead, phenomenology, and Merleau-Ponty himself, were
soon left behind, at least in France, by a new age dominated by
structuralism alone. Dosse2 gave two explanations for this ironical
situation: that Merleau-Ponty died too early, and that the structuralist
disciplines he advocated were far less concerned with his
phenomenological problems than he was with their methodology and
findings. To quote from Descombes in an interview by Dosse, those
disciplines «were already active in their own conceptual development
and did not need Merleau-Ponty or any other philosopher to interpret
their discoveries. They were all already at work on both levels».3
As the acknowledged “birthplace” as well as a prototypical domain
of structuralism, linguistics was undoubtedly among the disciplines
that Dosse and Descombes talked about. But curiously, there has not
been much discussion on the details of structuralist developments
within linguistics from the philosophical perspective – or from the
perspective of comparison between phenomenology and

1 BARTHES 1967, 24.

2 DOSSE 1997, 40.
3 DOSSE 1997, 40.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 91

structuralism. A historical reason Dosse gave for the minor role of

linguistics in his grand picture of how structuralism abandoned the
phenomenological subject and moved on in the human sciences is that
language studies, for institutional reasons, «remained very marginal»
in the French academia of the 1950’s.4 However, we are still tempted to
ask: how did the structural linguists, in their heyday, actually
understand structuralism as a guiding principle in their research, and
how did this understanding differ from that of Merleau-Ponty
philosophically? After all, Saussure, as Merleau-Ponty’s major source
of inspiration, was a linguist, and one of the few most important to
whom subsequent linguists keep returning. Therefore, a detailed
comparison of Merleau-Ponty’s Saussure and the linguists’ Saussure
will provide us with a direct access to where Merleau-Ponty and the
structuralists (in linguistics) differ, and thus an additional perspective
in the understanding of both phenomenology and structuralism.
In this study, I shall focus on how Merleau-Ponty and the structural
linguists interpreted Saussure’s differential theory. This theory, as is
well known, occupied the centre of Merleau-Ponty’s interest in
Saussure for the diacritical nature of the linguistic sign it claims. On
the other hand, it also constituted an important principle in linguistics
in the development of a series of fundamental notions and methods
including “phonemes” and “distinctive features”. Therefore, both
Merleau-Ponty and the linguists had much to say about the theory,
and, as we shall see, much versus each other.

2. “Difference” in Saussure’s linguistics

Before discussing Merleau-Ponty and the linguists’ interpretations of
Saussure’s differential theory, it is necessary that we take a look at the
notion of “difference” in Saussure’s linguistics first, since this is where
the theory originated, and the point of departure for all interpretations
to come.

4 DOSSE 1997, 66.

Metodo Vol. 4, n. 2 (2016)

 92 Lei Zhu

The normal way to approach Saussure’s “difference” is to go directly

to his Course in General Linguistics, a classic published posthumously
on the basis of students’ notes. But to appreciate the significance of
this notion in Saussure’s entire linguistics, one should really start with
Saussure’s earlier work. As Joseph has shown, even in the Essay for
Reducing the Words of Greek, Latin and German to a Few Roots, his first
effort in linguistics at the age of seventeen, Saussure demonstrated an
unusual approach to speech sounds more as distinctive units than as
actual pronunciations, which was «not characteristic of the linguistics
of Saussure’s time».5 Saussure argued that human languages
developed in several steps in which sounds multiplied and became
distinctive gradually. In the first step, there were only vowels; in the
second, the consonants k, p, and t emerged one by one and combined
with the vowels to form the syllables pa, ap, ka, ak, ta, and at, where a
was only an algebraic abstraction with no distinctive function between
[a], [o] or [e] as its realisation. In the third step, the syllables, through
the addition of a second consonant, multiplied into pap, pat, pak, tap,
tat, tak, kap, kat, and kak, each with a basic “root” meaning distin-
guishable by its special consonant pattern. It was only in the still next
steps that the vowels became distinctive for meaning, the consonants
developed their fricative or liquid variants, and multisyllabic roots
were formed.
The above theory, though rejected by Saussure himself later as a
result of youthful naïveté, obviously contained an insight on speech
sound and meaning that bears substantial similarity with the
diacritical view of linguistic signs that he was to develop. As Joseph
comments, Saussure’s vision here proved to be «fundamentally
consistent, whatever developments it would undergo, from his young
manhood to his death».6 Moreover, the priority of sound patterns over
actual pronunciations in the above theory was echoed later in
Saussure’s emphasis on the “sonant nasal” which, again, started in his
youthful years as a student and continued throughout his life.

5 JOSEPH 2012: 155

6 JOSEPH 2012, 158.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 93

Basically, the “sonant nasal” is a syllabic consonant used in the

reconstruction of Proto-Indo-European when reflexes of an etymon
have a vowel corresponding to a nasal consonant. For example, the
Greek root tetakh- (“to arrange”) should in theory take the suffix -ntai
in its third-person plural passive form, giving rise to tetakhntai, but
instead of this form, what has been found in literature is either a
periphrastic expression in its place or tetakhatai, with -ntai replaced by
-atai. Thus, to keep the overall sound pattern consistent, one has to go
beyond the normal distinction between vowel and consonant to
propose a “sonant nasal” in the protolanguage, which is both a
consonant and a syllable nucleus (like a vowel). Saussure noticed this
consonant-vowel exchange as a requirement of the sound pattern
when he was only sixteen. Although he was not the first to discover
the phenomenon (as he may have believed himself), it is certain that
for all his life, Saussure took it as an important enlightenment in his
encounter with language and linguistics.7
With the above as the background, we can now come to the Memoir
on the Original System of Vowels in the Indo-European Languages (Mémoire
sur le système primitif des voyelles dans les langues indo-européennes), for
which Saussure was most known in his lifetime. As we shall see, this
ingenious work is again characterized by the Saussurean respect for
overall sound patterns and disregard of individual pronunciations. A
problem that the work deals with concerns the irregularities in the
ablaut system (i.e., alternation of vowels as in English sing~sang~sung)
in Proto-Indo-European. While the regular ablaut of a Proto-Indo-
European root, like the Ancient Greek reflex for “leave” in the
leip~loip~lip alternation, usually involves three base forms, with one
containing an internal e (like leip, and thus called the “e-grade”),
another an internal o (like loip, and thus called the “o-grade”), and the
third one neither an internal e nor o (like lip, and thus called the “zero-
grade”),8 some Proto-Indo-European roots apparently show different
ablaut patterns in their reflexes. For example, the Greek root for

7 JOSEPH 2012, 132-5.

8 Following Brugmann, SAUSSURE [1878] 1922 used a1 and a2 to indicate the e and o here.

Metodo Vol. 4, n. 2 (2016)

 94 Lei Zhu

“stand” has stā~stō~sta instead of stea~stoa~sta, and the Greek root for
“give” has dō~dō~do instead of deo~dō~do. To explain these apparent
irregularities, Saussure9 proposed that the zero-grade forms of these
roots should have been *stA and *dOO in Proto-Indo-European,10 and
thus their ablaut forms *steA~*stoA~*stA and *deOO ~*doOO ~*dOO ,
following exactly the regular pattern. It was only later that *A, *OO , and
the combinations *eA, *oA, *eOO , and *oOO assumed their corresponding
pronunciations a, o, ā, ō, ō, and ō in Ancient Greek; they actually
followed different rules and assumed different sound values in
Sanskrit, Latin and other languages. Saussure never mentioned how
*A and *OO should have been pronounced in Proto-Indo-European; he
actually categorised them among the group of sounds he called
“sonant coefficients” (coefficients sonantiques), which included the
glides *i and *u, the nasals *mO and *n O (i.e., the “sonant nasals”
discussed above), and the liquid *rO 11. These “sonant coefficients” had
the same feature of showing more functional value than fixed
pronunciation: they could follow a vowel, like a consonant (as in the
“e-grade” and “o-grade” forms), but they could also serve as a syllable
nucleus, like a vowel (as in the “zero-grade” form).
Saussure’s solution was incomplete in that he should have proposed
an *E in addition to *A and *OO to account for such roots as the Greek
for “put”, whose ablaut pattern thē~thō~the would thus have derived
from the Proto-Indo-European *theE~*thoE~*thE. However, Saussure
certainly succeeded in establishing a completely new and ingenious
approach to the problem involved. Famously, Saussure’s *A was
shown (fourteen years after Saussure’s death) to have been attested in
the then newly-deciphered Hittite language, where traces of *A
showed up as a laryngeal.12 Since then, Saussure’s theory has been
known as the “laryngeal theory”, and the method that laid its
foundation has come to be called “internal reconstruction”. As we can

9 SAUSSURE 1922 [1878].

10 I follow the linguistic convention in this article to mark every reconstructed form with an
asterisk.
11 SAUSSURE 1922 [1878], 127.
12 KURYŁOWICZ 1927.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 95

see, essential to these contributions of Saussure – which are highly

acclaimed as «a paradigm shift in the Kuhnian sense» 13 – is a shift of
focus from the changes of concrete sounds by themselves to patterns
internal to a language in which functional units differ from each other
and apparent irregularities are merely results of the interactions of the
units. Had Saussure dwelt on the individual histories of the Greek stā,
stō, sta and so on rather than the general patterns of vowel alternation,
he would never have come up with the hypothetical *A and *OO .
Actually, his concern with patterns was such that he cared little for the
sound values of *A and *OO as long as these elements fit into the
patterns where their functional values remained constant.
Thus, as we finally come to Saussure’s Course in General Linguistics,
the famous claim that «in a language there are only differences, and no
positive terms»,14 which stands among the most original contributions
of the work, is hardly surprising. It is a statement of a guiding
principle in Saussure’s previous work, with extensions and
elaborations on two important levels: the synchronic and the
semiological.
In Saussure’s previous work, especially the Memoir on the Original
System of Vowels in the Indo-European Languages, the focus was placed
on the historical changes of language and the genealogical relations
they reveal. This was natural in Saussure’s time, but Saussure’s
approach to these historical issues was already rather “ahistorical” in
the sense that he engaged himself far less with the histories of
individual elements of a language than with the elements’ patterning,
or interrelations. For him, individual sound values were not important
by themselves; if constant interrelations could be observed, he would
rather venture the unconventional by postulating elements which
appeared as unusual or unknown sounds, but kept the interrelations
intact. The “sonant nasal” he advocated, for instance, was an unusual
sound that maintained the relationship “a=n”; the *A and *OO he
hypothesised, further, were unknown sounds that helped to preserve

13 CLACKSON 2007, 53.

14 SAUSSURE 1983, 118.

Metodo Vol. 4, n. 2 (2016)

 96 Lei Zhu

the interrelations in the Indo-European ablaut, and whose actual

pronunciations he did not even bother to discuss. Therefore,
individual sounds to Saussure really meant little more than slots in a
web of interrelations, and their histories the superficial manifestations
of the web working underneath. In the Course in General Linguistics,
Saussure expressed the idea explicitly:

The diachronic developments are in no way directed

towards providing a new sign to mark a given value. […]
These diachronic events do not even tend to change the
system. […] the language system as such is never directly
altered. It is in itself unchangeable. Only certain elements
change, but without regard to the connexions which
integrate them as part of the whole.15

What is stated here can actually be regarded as a summary of an

important principle in Saussure’s own historical linguistic work. At the
same time, it also signals a natural transition from diachrony to
synchrony, since the “unchangeable system” is now clearly brought up
as the more essential aspect of language despite the changes of its
elements observed on the surface. And such a system, without the
presumed status of individual elements, is nothing but a web of
interrelations, or differences.
The notion of “difference” was formally brought to the fore in
Saussure’s further discussion of the semiological principles with
which he defined the “unchangeability” of the linguistic system. As
Saussure saw it, the system is unchangeable not because the actual
interrelations in a language are closed to variation, but because
intrinsic to all linguistic systems are «the unchanging principles of
semiology»,16 whose essence he later summarised thus:

Everything we have said so far comes down to this. In the

15 SAUSSURE 1983, 84.

16 SAUSSURE 1983, 88.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 97

language itself, there are only differences. Even more important

than that is the fact that, although in general a difference
presupposes positive terms between which the difference
holds, in a language there are only differences, and no
positive terms. Whether we take the signification or the
signal, the language includes neither ideas nor sounds
existing prior to the linguistic system, but only conceptual
and phonetic differences arising out of that system.17

Compared with what we can infer from Saussure’s practice in his

historical linguistic studies, the statement above presents two obvious
developments in Saussure’s thought: (1) For the first time, Saussure
overtly declared the lack of ontological status of individual sounds, a
principle that had underlain his linguistic studies since his youthful
days. (2) Generalising about the linguistic sign, Saussure extended his
differential view of the speech sound to signification, or concept/idea.
As he repeatedly pointed out,

If words had the job of representing concepts fixed in

advance, one would be able to find exact equivalents for
them as between one language and another. But this is not
the case. […] In all these cases what we find, instead of
ideas given in advance, are values emanating from a
linguistic system. If we say that these values correspond to
certain concepts, it must be understood that the concepts in
question are purely differential. […] But it must not be
supposed that the concept in question has any kind of
priority. On the contrary, that particular concept is simply a
value which emerges from relations with other values of a
similar kind.18

It is statements like the above, which bear directly upon meaning

17 SAUSSURE 1983, 118.

18 SAUSSURE 1983, 115-6.

Metodo Vol. 4, n. 2 (2016)

 98 Lei Zhu

and resolutely repudiate any pre-given concept, that have attracted the
widest attention among Saussure’s readers. Having established the
differential view on both speech sounds and concepts, Saussure
further united it with the arbitrary nature of language, so that his
semiology accounted for not only the “signifier” (signifiant, i.e., the
sound, or sound image, as distinct from actual pronunciation) and the
“signified” (signifié, i.e., the concept), but also their relations. As both
sound and concept are differentially determined instead of fixed in
advance, it is natural that as long as the differential values are kept, no
concept must be related to any specific sound. Hence Saussure’s
assertion that «the terms arbitrary and differential designate two
correlative properties».19

3. Merleau-Ponty’s interpretation of the differential

theory
Saussure’s theory about the differential nature of the linguistic sign,
growing out of his historical linguistic practice and ripening in his
Course in General Linguistics, was not meant so much to address the
circle of philosophers as to urge a revolution in linguistics regarding
its object of study and methodology. However, its “discovery” and
dissemination by Merleau-Ponty changed the situation to a great
extent. As is generally acknowledged, Merleau-Ponty did not begin to
read Saussure seriously until the late 1940’s, 20 but was soon attracted
by Saussure’s ideas once he did, and highly commended them as a
major source of inspiration until his premature death in 1961. It has
also been pointed out that of all the Saussurean notions that Merleau-

19 SAUSSURE 1983, 116.

20 Merleau-Ponty may have been aware of Saussure’s work some time earlier, but
Saussure’s name first appeared in his writings in 1947, in the article The Metaphysical in
Man. Extensive discussions of Saussure appeared still later, in the works composed in the
early 1950’s, including the article Indirect Language and the Voices of Silence and the
unfinished The Prose of the World. See SCHMIDt 1985: 105; 194 and KAGANOI 1998: 152-55 for
details.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 99

Ponty drew upon, “difference” was undoubtedly the central one, and
almost the only one that Merleau-Ponty did not “misunderstand” or
alter significantly.21 As Kaganoi says,

Then what, strictly speaking, did Merleau-Ponty discover in

Saussure’s theory? It cannot be emphasized too strongly
that what he discovered and thoroughly understood is only,
in Saussure’s words, the “differential theory of signs”
which is summarized at the beginning of “Indirect
Language and the Voices of Silence”. At this point, some
may claim to say that he also adopted some other notions
such as “diachrony”, “synchrony”, “signifier”, “signified”,
etc. but almost all of these other terms Merleau-Ponty used
in a way against Saussure’s intentions.22

It was not by accident, however, that Merleau-Ponty got the

differential theory “right” and the other notions “wrong”. Kaganoi
himself observed from Merleau-Ponty’s “correct” account of these
notions in his lecture notes and his insistence on the “incorrect”
version elsewhere that the “incorrect” version «was not due to bad
memory or misunderstanding, but to his own intentional
reinterpretation of Saussure».23 The same, of course, can be said of
Merleau-Ponty’s “correct” version of the differential theory as well,
since it was no less a part of the reinterpretation, and indeed the one
occupying the central place. I will show in this section that the
designation of Merleau-Ponty’s Saussure as “correct” in one place and
“incorrect” in another is actually not very helpful, because the
differential theory also underwent significant transformations by
Merleau-Ponty, and has to be understood against Merleau-Ponty’s
general philosophical endeavor just as the other Saussurean ideas
involved.

21 LAGUEUX 1965; RICOEUR 1967; SCHMIDT 1985.

22 KAGANOI 1998, 155.
23 KAGANOI 1998, 156.

Metodo Vol. 4, n. 2 (2016)

100 Lei Zhu

There have been many discussions on why Saussure’s differential

theory should be so important to Merleau-Ponty – a basic question for
all who intend to trace Merleau-Ponty’s thought after the
Phenomenology of Perception. While various complexities are involved,
what has proved to be the most significant to this question are the two
fundamental stances in Merleau-Ponty’s thought, which Gutting 24 has
summarised as his emphasis on overcoming «the common-sense
prejudices associated with the ‘natural attitude’ in which we
unreflectively encounter the world» and his objection to achieving this
goal by «a return to a pre-personal transcendental subject before
which the entire world appears with complete transparency». The first
of these is a basic phenomenological stance, the overcoming of the
“natural attitude” – originally a term from Husserl – being Merleau-
Ponty’s favorite way to explain phenomenology, as one can see from
the following quotations from his first and last writings:

[…] Not because we reject the certainties of common sense

and a natural attitude to things […] but because, being the
presupposed basis of any thought, they are taken for
granted, and go unnoticed, and because in order to arouse
them and bring them to view, we have to suspend for a
moment our recognition of them.25

We see the things themselves, the world is what we see:

formulae of this kind express a faith common to the natural
man and the philosopher […] [Ceaselessly the philosopher
finds himself] obliged to reinspect and redefine the most
well-grounded notions, to create new ones […] at whose
term the evidence of the world […] is supported by the
seemingly most sophisticated thoughts, before which the

24 GUTTING 2001, 187.

25 MERLEAU-PONTY 1962, xiv-xv.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 101

natural man now no longer recognizes where he stood.26

The second stance, however, is targeted at a tendency in

phenomenology, as the philosopher suspends our notions of the
world, to reduce the notions to individual perceptions or pure
experience, thus introducing an absolute soul as the ultimate
guarantee, and going against the original reflective attitude of
phenomenology. This stance resulted from Merleau-Ponty’s acute
awareness of the delicate position of phenomenology between
intellectualism and empiricism, which he made clear at the beginning
of his preface to the Phenomenology of Perception:

It [phenomenology] is a transcendental philosophy which

places in abeyance the assertions arising out of the natural
attitude, the better to understand them; but it is also a
philosophy for which the world is always 'already there'
before reflection begins – as an inalienable presence […] It
is the search for a philosophy which shall be a 'rigorous
science', but it also offers an account of space, time and the
world as we 'live' them.27

In the same book, Merleau-Ponty further elaborated on the point

that intellectualism and empiricism are antitheses «on the same
level»28 where the subjective and the objective are opposed to each
other in principle, and the world is reduced to an external cause, be it
the scientific reason or the transcendental subject. Merleau-Ponty was
especially cautious, in the discussion of perception, not to ground our
bodily experience upon a Cartesian mind (cogito) independent of the
perceived. At the same time, meaning and certainty being undeniable,
he proposed a «tacit cogito, the presence of oneself to oneself», which
exists before and “conditions” the spoken language, but is not

26 MERLEAU-PONTY 1968, 3.
27 MERLEAU-PONTY 1962, vii.
28 MERLEAU-PONTY 1962, 23.

Metodo Vol. 4, n. 2 (2016)

102 Lei Zhu

transcendental, since it «has upon itself and upon the world only a
precarious hold», «like that of the infant at its first breath, or of the
man about to drown and who is impelled towards life».29
Merleau-Ponty’s effort to keep this delicate position while pursuing
his phenomenological analysis of perception, however, was not
appreciated by everyone. As is indicated by the minutes of a
conference shortly after the publication of the Phenomenology of
Perception, many people regarded Merleau-Ponty’s idea as an
unacceptable grounding of reason upon subjectivity and
inexactitude.30 Judging from his own rejection of the tacit cogito in his
last work The Visible and the Invisible, Merleau-Ponty did not stop
thinking about the problem himself. Therefore, it is not difficult to
imagine his excitement upon the encounter with Saussure’s
differential theory of the linguistic sign: through the denial of all pre-
given concepts, the theory provided a way out of the “natural
attitude”; at the same time, drawing on “structure” – or “system”, to
use Saussure’s term – as the place of origin where concepts emanate
through relations with each other, it avoided the idealist inclination in
phenomenology towards subjectivity.
In The Prose of the World, Merleau-Ponty explicitly pointed out that
although the “natural attitude” prevailed in the exact sciences, where
words were taken without reflection as expressions of essences of the
world, linguistics was an exception:

[…] it is thought that language (like God’s understanding)

contains the germs of every conceivable signification and
that all our thoughts are destined to be expressed in
language; or that every signification which enters man’s
experience carries within it its own formula, as the sun, in
the minds of Piaget’s children, bears its name in its center.
Our language recovers, in the heart of things, the word
which made the thing.

29 MERLEAU-PONTY 1962, 470.

30 MERLEAU-PONTY 1964b.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 103

These are not just common-sense attitudes. They predo-

minate in the exact sciences (but not, as we shall see, in
linguistics).31

The findings of the sciences of language are decisive here.

[…] Today psychology and linguistics reveal that it is in fact
possible to forsake a timeless philosophy without falling
into irrationalism. Saussure shows admirably that […] it
cannot be the history of the word or language which
determines its present meaning; […].32

In the ensuing discussion to what has been quoted here, Merleau-

Ponty made it clear that the Saussurean idea he was talking about was
exactly the differential theory of the linguistic sign. Without doubt,
when describing the theory as a means to «forsake a timeless
philosophy without falling into irrationalism», Merleau-Ponty was not
only referring to its value in overcoming the predominant “natural
attitude”, but was thinking of its special advantage in overcoming the
attitude without incurring an idealist tendency or the kind of mis-
understanding as he had to face after the publication of the
Phenomenology of Perception. Actually, the tacit cogito which Merleau-
Ponty proposed in the Phenomenology of Perception, as a silent pre-
conditioning of the spoken language, resembles the language-
generating “structure” in significant ways, but the “structure”, which
gives rise to language through pure relations, is far more removed
from cogito and from individual subjectivity. In The Prose of the World,
Merleau-Ponty no longer appealed to the tacit cogito; with the differen-
tial theory, his presentation of the mechanism in the forming of words
in language became more specific, elaborate, and resistant to the
temptation of subjectivity:

31 MERLEAU-PONTY 1973, 4.
32 MERLEAU-PONTY 1973, 22.

Metodo Vol. 4, n. 2 (2016)

104 Lei Zhu

[…] the words and very forms for an analysis of this kind
soon appear to be secondary realities, the results of a more
originary differentiation. The syllables and letters, the turns
of phrase, and the word endings are the sediments of a
primary differentiation […] The phonemes, too, which are
the real foundations of speech […] They bring us into the
presence of that primary operation, beneath
institutionalized language, that creates the simultaneous
possibility of significations and discrete signs.33

In the “more originary differentiation”, which settles down into

phonemes and syllables and forms the foundation of language,
Merleau-Ponty saw the prospect of a philosophy that would serve his
reflective purpose with a clearer theory about the origin of concept
while maintaining the delicate non-intellectualist and non-empiricist
position that he had emphasised. In the article From Mauss to Claude
Lévi-Strauss, Merleau-Ponty referred to such a philosophy specifically
when commenting on Lévi-Strauss’s structuralist anthropology:

This notion of structure, whose present good fortune in all

domains responds to an intellectual need, establishes a
whole system of thought. For the philosopher, the presence
of structure outside us in natural and social systems and
within us as a symbolic function points to a way beyond the
subject-object correlation which has dominated philosophy
from Descartes to Hegel.34

Thus, on the basis of his own philosophical considerations, Merleau-

Ponty accorded to Saussure’s differential theory an ontological
significance not entertained by Saussure himself. Originally an
argument in support of the priority of synchrony over diachrony in the
study of language, the differential theory was now transformed into a

33 MERLEAU-PONTY 1973, 32-3.

34 MERLEAU-PONTY 1964a, 123.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 105

new philosophy that promised to break away from the intellectualism-

empiricism dichotomy.
At the same time, Merleau-Ponty did not really disregard Saussure’s
concern with synchrony and diachrony; on the contrary, he tried
further to develop, along this line, a new interpretation of history from
the differential theory and its philosophical implications. In his inau-
gural address at the Collège de France (1953), Merleau-Ponty argued
that «[t]he theory of signs, as developed in linguistics, perhaps implies
a conception of historical meaning which gets beyond the opposition
of things versus consciousness» and that «Saussure could well have
sketched out a new philosophy of history».35 Such a declaration should
not be surprising if one returns to The Prose of the World, which
Merleau-Ponty had been writing during the two years before the
address. There, the introduction of Saussure’s differential theory actu-
ally started with an intensive discussion on how words are first and
foremost results of a primary differentiation in structure rather than
individual existences with their own disparate histories – i.e., etymo-
logies. In a sense, this view was a continuation of Saussure’s principle
in his historical linguistic reconstruction which, as mentioned in the
previous section, prioritised sound patterns over individual pronun-
ciations in revealing the essential mechanisms of language change: the
correspondence between the reflexes of the “sonant nasals”, for
instance, was much more important than the reflexes by themselves;
and the value of the “sonant coefficients” in maintaining the Indo-
European ablaut system far exceeds the sound changes of these
“sonant coefficients”. But in The Prose of the World, Merleau-Ponty, fol-
lowing the Saussure of the Course in General Linguistics, extended the
historical priority of structure over individuality beyond the world of
speech sounds. All the changes in the history of language, individual
incidents as they appear to be, should now be viewed as «incorporated
into a self-maintaining system of expression» through «coherent
motivations».36 On the other hand, this “self-maintaining system” is

35 MERLEAU-PONTY 1963, 54-5.

36 MERLEAU-PONTY 1973, 24.

Metodo Vol. 4, n. 2 (2016)

106 Lei Zhu

not a new fixed object, for the observed categories that it generates in
this or that language are but «a retrospective and inessential expres-
sion of our own power of speech». 37 The system is neither totally
personal nor totally outside the person, since it is «what most belongs
to us as individuals», and «while addressing itself to others, it simul-
taneously acquires a universal value».38 Thus, the history of language
is neither a collection of discrete incidents nor the expression of an
absolute and objectified being. It was in this sense that Merleau-Ponty
credited Saussure with the conception of a new philosophy of history:

In any case, Saussure has the great merit of having taken

the step which liberates history from historicism and makes
a new conception of reason possible. […] Objective history
was – as is all history, for Saussure – an analysis which
decomposes language and in general institutions and
societies into an infinite number of accidents.39

The significance of the new philosophy of history to “institutions

and societies” in general was then Merleau-Ponty’s central concern
when he referred to this new philosophy in his inaugural address at
the Collège de France. As he says, «[t]he presence of the individual in
the institution, and of the institution in the individual is evident in the
case of linguistic change».40 By the same token, events and individuals
are to be understood through their values in a system that is not totally
outside them at the same time.
As Culler41 has observed, linguistics, with Saussure’s theory about
the linguistic sign, is a model for anyone who intends to «avoid the
familiar mistake of assuming that signs which appear natural to those
who use them have an intrinsic meaning and require no explanation».
In this respect, Merleau-Ponty was indeed one of the first to recognise

37 MERLEAU-PONTY 1973, 26.

38 MERLEAU-PONTY 1973, 85.
39 MERLEAU-PONTY 1973, 23-4.
40 MERLEAU-PONTY 1963, 55.
41 CULLER 1975, 6.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 107

the philosophical significance of Saussure. However, Merleau-Ponty’s

recognition had a clear phenomenological motivation. By now, we
have seen how Merleau-Ponty transformed Saussure’s differential
theory into a new ontology as well as the foundation of a new philo-
sophy of history. These transformations show evidently that Merleau-
Ponty was much more than a transmitter of the differential theory. In a
sense, the differential theory for Merleau-Ponty was only “raw
material” to be incorporated into his own framework. He not only
built upon it, but trimmed and altered it from his own perspective.
An example of Merleau-Ponty’s trimming of the differential theory is
his restraint on the use of the terms “signifier” and “signified” in the
Saussurean sense when discussing the differential nature of the two
sides of the sign. As Kaganoi42 has pointed out, Merleau-Ponty, being
acutely aware of the danger of looking at language “objectively”,
would rather keep referring to “phonetic differences” and “conceptual
differences” than consider the “signifier” and the “signified”
separately as “acoustic image” and “concept” as Saussure did. In the
few places where he did talk about the “signifier” and the “signified”,
his remarks seem very peculiar, e.g.:

Let us say that every expression is perfect to the extent it is

unequivocally understood, and admit as a fundamental fact
of expression a surpassing of the signifying [i.e. signifier]
by the signified which it is the very virtue of the signifying
to make possible.43

What Merleau-Ponty intends to convey here is actually the idea that

the “signified”, rather than being pre-fixed, derives from an originary
differentiation, and that once it is derived, it establishes a way to talk
about the world with a structure in which everything occupies a place
different from the others. Thus, it is through language that our
understanding of the world is established: we go from the “signifier”

42 KAGANOI 1998, 157.

43 MERLEAU-PONTY 1964a, 90.

Metodo Vol. 4, n. 2 (2016)

108 Lei Zhu

to a world of ideas, which seem natural as a “perfect” way to present

everything, but precisely because of the disclosure of this process, we
know at the same time that the ideas are not pre-fixed, but derived.
Here, we start to see the divergence between the linguist and the
phenomenologist: although the “signifier” and the “signified” are
perfectly legitimate terms for the linguist, the phenomenological
philosopher perceives a tendency in these names to substitute secon-
dary entities for their originary sources. While Merleau-Ponty was
dissatisfied with the idealist inclination in phenomenology towards
subjectivity, we can sense here that he was equally on guard against
the tendency towards objectivity. To prevent language from being
taken as a static combination of objective entities, he tried to highlight
the lively process of the derivation of concepts from differentiation.
Merleau-Ponty’s caution against objectivity here has a larger
background. For he knew very well that the area of linguistics, where
the differential theory had originally been proposed, was after all an
area of science, just as the other human sciences that followed up in
adopting the structuralist paradigm. As early as in the Phenomenology
of Perception, Merleau-Ponty declared:

I cannot conceive myself as nothing but a bit of the world, a

mere object of biological, psychological or sociological
investigation. I cannot shut myself up within the realm of
science.44

Therefore, while holding up the differential theory as an important

breakthrough in his later philosophy, Merleau-Ponty could not but
remind his reader that the newly-found significance in “structure”
should not be misunderstood as a retreat to a Platonic idea. When
commenting on Lévi-Strauss’s structuralism in anthropology, he said:

[…] when the scientist formulates and conceptually

determines structures, and constructs models by means of

44 MERLEAU-PONTY 1962, ix.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 109

which he intends to understand existing societies, there is

no question for him of substituting the model for reality. As
a matter of principle, structure is no Platonic idea. To
imagine imperishable archetypes which dominate the life
of all possible societies would be to make the mistake the
old linguistics made when it supposed that there was a
natural affinity for a given meaning in certain sonorous
material.45

But the problem is: if “structure” is not a Platonic idea, can we ask,
from the phenomenological point of view, how the actual structures in
particular languages or societies actually take form on the basis of the
originary differentiation? This takes us back to a basic phenomeno-
logical task Merleau-Ponty raised in the Phenomenology of Perception
concerning the foundation of science:

The whole universe of science is built upon the world as

directly experienced, and if we want to subject science itself
to rigorous scrutiny and arrive at a precise assessment of its
meaning and scope, we must begin by reawakening the
basic experience of the world of which science is the
second-order expression.46

So, the question now is: if structure tends to be objectified in science

as this or that particular structure, can the phenomenologist
“reawaken” the world of the originary differentiation of which these
objectified structures are but “second-order” expressions? Or, to be
more specific, if there are only differences in the unobjectified
structure, how do particular concepts or meanings arise in the first
place? These questions are actually already there in the often-quoted
beginning lines of Merleau-Ponty’s article Indirect Language and the
Voices of Silence:

45 MERLEAU-PONTY 1964a, 115.

46 MERLEAU-PONTY 1962, ix.

Metodo Vol. 4, n. 2 (2016)

110 Lei Zhu

What we have learned from Saussure is that, taken singly,

signs do not signify anything, and that each one of them
does not so much express a meaning as mark a divergence
of meaning between itself and other signs. Since the same
can be said for all other signs, we may conclude that
language is made of differences without terms; or more
exactly, that the terms of language are engendered only by
the differences which appear among them. This is a
difficult idea, because common sense tells us that if term A
and term B do not have any meaning at all, it is hard to see
how there could be a difference of meaning between them;
and that if communication really did go from the whole of
the speaker’s language to the whole of the hearer’s
language, one would have to know the language in order to
learn it. But the objection is of the same kind as Zeno’s
paradoxes; and as they are overcome by the act of
movement, it is overcome by the use of speech. And this
sort of circle, according to which language, in the presence
of those who are learning it, precedes itself, and suggests its
own deciphering, is perhaps the marvel which defines
language.47

Here we can see clearly that Merleau-Ponty, after all, was a

phenomenologist. Although he adopted the differential theory for its
advantages in helping him articulate and develop his philosophical
system, he never forgot the phenomenological task he set out for
himself in the Phenomenology of Perception. His question «how there
could be a difference of meaning» between items that «do not have any
meaning at all» is a typical phenomenological one in the line of
Husserl’s genealogy of ideas which Saussure did not consider or
answer. Merleau-Ponty’s own treatment of the question here is rather
brief, but drawing on Zeno’s paradoxes, he did present us some key
notions in his approach to the “phenomenology of structure” – the

47 MERLEAU-PONTY 1964a, 39.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 111

“use of speech” in particular – which actually defined the character of

his later philosophy.
To reach a comprehensive understanding of the phenomenological
question about differentiation and the forming of concepts, Merleau-
Ponty reached back to his earlier arguments in the Phenomenology of
Perception, trying to combine some of them with Saussure’s theory. In
the Phenomenology of Perception, for example, Merleau-Ponty proposed
a distinction between the “speaking word” (parole parlante) and the
“spoken word” (parole parlée) on the basis of the “celebrated
distinction” between “languages” (langages) and “speech” (parole): just
as languages are «both the repository and residue of acts of speech, in
which unformulated significance not only finds the means of being
conveyed outwardly, but moreover acquires existence for itself, and is
genuinely created as significance», the “spoken word” is the
completed state of the “speaking word” where «the significant
intention is at the stage of coming into being». 48 While the distinction
here between langages and parole was not the same as Saussure’s
langue-parole distinction even in terminology,49 Merleau-Ponty made
the following claim in The Prose of the World:

Alongside the linguistics of language, which gives the

impression, in the extreme, that language is a series of
chaotic events, Saussure has inaugurated a linguistics of
speech, which would reveal in it at each moment an order, a
system, a totality without which communication and the
linguistic community would be impossible.50

This claim has long been criticised as an example of Merleau-Ponty’s

idiosyncratic rendition of Saussure’s theory against the latter’s
intention. But from the context, one can see that the “linguistics of
language” here is in fact a reference to the linguistics before Saussure,

48 MERLEAU-PONTY 1962, 229.

49 For Saussure (1983, 77), langage was the general term used to refer to langage plus parole.
50 MERLEAU-PONTY 1973, 23.

Metodo Vol. 4, n. 2 (2016)

112 Lei Zhu

which focused on the history of individual sounds and words,

whereas the “linguistics of speech” is a general reference to Saussure’s
own linguistics, with the differential theory as a principle. The crux of
the claim here, then, is Merleau-Ponty’s use of “speech” as opposed to
“language” in his references. This shows that for Merleau-Ponty, the
originary differentiation, which he saw as Saussure’s major
contribution to disclose, was more relevant to the act of speaking than
to its static residue. In the Prose of the World, Merleau-Ponty continued
to talk about the “speaking language” (langage parlant) and the
“spoken language” (langage parlé) and to use examples from the child’s
first experience of language acquisition and the writer or reader’s
experience with a new style of writing – as he did in the
Phenomenology of Perception – to illustrate how the first meanings are
generated in the act of speaking. In the article Indirect Language and the
Voices of Silence, Merleau-Ponty further elaborated on how the child’s
first perceptions of phonemic differences in speech gradually lead to
the establishment of a structure where concepts are formed:

It can be said that beginning with the first phonemic oppo-

sitions the child speaks, and that thereafter he will only
learn to apply the principle of speech in diverse ways.
Saussure’s insight becomes more precise: with the first pho-
nemic oppositions the child is initiated to the lateral liaison
of sign to sign as the foundation of an ultimate relation of
sign to sense – in the special form it has received in the lan-
guage in question. […] The untiring way in which the train
of words crosses and recrosses itself, and the emergence
one unimpeachable day of a certain phonemic scale
according to which discourse is visibly composed, finally
sways the child over to the side of those who speak.51

Thus, by incorporating Saussure’s ideas into his own earlier

discussion on how meaning is generated in speaking and settles down

51 MERLEAU-PONTY 1964a, 40-41.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 113

as sediments in language, Merleau-Ponty brought the differential

theory back to the track of his phenomenological exploration. To give
an explanation to the actual generation of structures through
differentiation, he highly emphasised the significance of the lively
process of speaking. This is why “speech” for him had a more
primordial sense than “language”, and this led to a series of
differences between him and Saussure. For Saussure, language was
more essential than speech, the latter being the realisation of the
former under concrete circumstances; but for Merleau-Ponty,
language, as the static sediments of a live system where there are only
differences, must be rooted in something more fundamental – and this
he found in speech. A relevant issue here is Merleau-Ponty’s attitude
to the “arbitrariness” of the sign, which Saussure regarded as
correlative with the sign’s differential nature. According to Merleau-
Ponty,52 the “original form” of any language is «not entirely arbitrary».
As he famously said in the Phenomenology of Perception:

[…] the words, vowels and phonemes are so many ways of

“singing” the world, and that their function is to represent
things not, as the naïve onomatopoeic theory had it, by
reason of an objective resemblance, but because they
extract, and literally express, their emotional essence.53

We do not find such statements in Merleau-Ponty’s later philosophy,

probably because of the inclination for subjectivity they seem to
suggest. But understanding that even after accepting Saussure’s
differential theory, Merleau-Ponty was still mainly interested in how
meaning is generated in the originary differentiation of speech instead
of the institutionalised language, we should not expect him to change
his basic attitude and subscribe to Saussure’s arbitrariness principle.
“Arbitrary” is how we see language when it is already a fixed system
in which relations are there as sediments of differentiation, but

52 MERLEAU-PONTY 1962, 217.

53 MERLEAU-PONTY 1962, 217.

Metodo Vol. 4, n. 2 (2016)

114 Lei Zhu

Merleau-Ponty was more interested in the process of differentiation

itself and how it led to the first meaningful expression.
On Merleau-Ponty’s intention in advocating structuralism as a
phenomenologist, Dosse has made the following remark:

[H]e [i.e., Merleau-Ponty] believed that the phenomeno-

logical philosopher should appropriate the works of Mauss,
Lévi-Strauss, Saussure, and Freud, not so much in order to
provide epistemological bases for each of these disciplines
as to subject them to a thoroughgoing phenomenological
renewal, which would redefine them philosophically,
assuming of course that the philosopher accepted the
validity of the specialists’ information, which he could not
verify. The phenomenologist was like an orchestra leader
drawing together all the objective results produced by the
social sciences and assigning them a meaning, a value in
terms of subjective experience and of total meaning.54

In this section, we have seen Merleau-Ponty’s interpretation of

Saussure’s differential theory exactly as Dosse has described. The
differential theory attracted Merleau-Ponty precisely because Merleau-
Ponty, from his own stance, saw the advantage of the theory in
achieving the phenomenological goal of reflection without the
inclination for subjectivity that he found difficult to escape. He
incorporated the theory into his own framework, and transformed it
into a new ontology as well as the foundation of a new philosophy of
history. At the same time, to counteract the theory’s objective
inclination and make up for its inability to explain how concepts are
derived from pure “differences”, Merleau-Ponty altered Saussure’s
idea significantly, making “speech” the primordial state where
meaning is generated through the originary differentiation.55
54 DOSSE 1997, 39.
55 In this section, I have not touched upon the idea of écart (divergence) in Merleau-Ponty’s
later philosophy, partly because the idea is much more than an interpretation of the
differential theory, and partly because it is so vague that its elucidation will require

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 115

4. “Difference” in post-Saussurean linguistics

It is generally accepted in the history of linguistics that «the most
immediate and historically one of the most important effects of de
Saussure’s structural theory of language was in the realm of
phonology», which then provided «[b]y far the strongest impulse to a
revision of theory and of its associated concepts» in the other branches
of linguistics.56 In this sense, Saussure’s differential theory is of utmost
importance to linguistics as a whole, since although the term
“phoneme” was not really Saussure’s own invention and meant
something different in Saussure’s terminology,57 the differential theory
provided the most articulate exposition of the principle underlying the
standard definition of “phoneme”, and thus laid the foundation for
phonology and for a series of major developments in linguistics. As
Saussure says:

The sound of a word is not in itself important, but the

phonetic contrasts which allow us to distinguish that word
from any other. That is what carries the meaning.58

much more space than this paper can afford. However, it should be pointed out that, as
HASS 2008: 173 says, «Saussure’s notion of ‘diacritical’ difference is the direct inspiration
for écart in The Visible and the Invisible». At the same time, one should also be aware that
écart, within the philosophy of the flesh, has a rather different meaning from that of the
Saussurean differences between concepts, phonemes or signs. In the Heideggerian sense,
it is «the absolute divergence between the being (which is blind to its own sense of
Being) and the unconcealment of that by which the being is» (MALDINEY 2000: 67).
56 ROBINS 1997, 226; 238.
57 As a word, “phoneme”, or its French origin “phonème”, was «an adaptation of an
attested Ancient Greek word, φώνημα (phōnēma)» which meant “voice”. It first
appeared in a French work by the Bulgarian philosopher Petăr Beron, and was later
adopted by the French amateur scholar Antoni Dufriche-Desgenettes to mean actual
pronunciation (MUGDAN 2011). Such was also what it meant in Saussure’s works. The
first to use the term in its standard modern linguistic sense was the Polish linguist
Mikołaj Kruszewski, a student of Jan Baudouin de Courtenay who is known for having
proposed the idea of “phoneme” at around the same time as Saussure.
58 SAUSSURE 1983, 116.

Metodo Vol. 4, n. 2 (2016)

116 Lei Zhu

Thus, for instance, if [p] and [ph] are used to distinguish one word
from another in Chinese (e.g., [pi35] “nose” versus [phi35] “skin”), but
not so in English, then they are two phonemes in Chinese, but one and
the same phoneme in English.
Apparently, the idea of the phoneme, together with Saussure’s
differential theory as its underlying philosophy, started to gain wide
currency among linguists from the 1910’s onward. 59 However, in the
application of the theory to the actual analysis of phonemes in
particular languages, the theory started to change. On the one hand,
these changes involved the transformation of the theory into concrete
and practical criteria to guide the actual analysis. For example, the
criterion of “complementary distribution” stipulates that basically, if
two sounds never occur in the same phonetic surroundings, they
should be regarded as belonging to the same phoneme. But on the
other hand, many phonologists started to add criteria that appeared to
be necessary complements to the differential theory. Of these, the most
important was the criterion that sounds must be similar to be grouped
into one phoneme. In his famous article The Non-uniqueness of
Phonemic Solutions to Phonetic Systems (1934), Chao summarised the
criterion as «phonetic accuracy, or smallness of range of phonemes»,
and argued for it thus:

A minimum degree of phonetic accuracy is provided for by

the 'similar in character' clause contained in Jones’s later
definition. By our purely logical definition, we should have
the possibility of regarding English [h] and [ŋ] as members
of one phoneme, which never occur in the same phonetic
environment, and we could write forms like [ɧæt], [biˈɧeiv],
[sɔɧ], [ˈsiɧə] for hat, behave, song, singer, and learn very
quickly when to say [h] and when to say [ŋ]. Such practice,
however, would not be favored by either the phonetician or
the philologist. […] Since, therefore, the automaticity of
variation is mostly of conditional nature, we shall have to
59 ROBINS 1997, 233.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 117

allow a good deal of latitude in the interpretation of the

'similar in character' clause.60

According to Chao, simply saying that «sounds must be similar to be

grouped into one phoneme» is not enough, because there is no
universal standard for phonetic similarity that is accepted by speakers
of all languages. However, despite the difference between “phonetic
similarity” and Chao’s more careful wording, the fundamental
concern here with concrete pronunciation and the rejection of the
“purely logical definition” of the phoneme on the basis of such a
concern are by no means what one can derive from Saussure’s differ-
ential theory. As is clear from our previous discussion, Saussure was
able to make his breakthrough in the reconstruction of Indo-European
laryngeals precisely because he insisted on the priority of sound
patterns over actual pronunciation. Had he put the criterion of
phonetic similarity or accuracy before the “purely logical” sound
patterns, he would not have come up with the “sonant nasal” solution
to the correspondence between Greek a and n in the first place.
Therefore, the above criterion from Chao actually stands in sharp
contrast with Saussure’s original proposal. However, Chao’s idea was
shared by most of his contemporaries. Moreover, in addition to the
above criterion, Chao proposed six other criteria, all of which have
been accepted as basic principles in phonemic analysis:

[…] (b) simplicity or symmetry of phonetic pattern for the

whole language, (c) parsimony in the total number of
phonemes, (d) regard for the feeling of the native speaker,
(e) regard for etymology, (f) mutual exclusiveness between
phonemes, (g) symbolic reversibility […].61

Of these criteria, some are more congruent with the differential

theory (e.g., (f)), while some, like the one about “phonetic accuracy”,

60 CHAO 2006 [1934], 250-1.

61 CHAO 2006 [1934], 250.

Metodo Vol. 4, n. 2 (2016)

118 Lei Zhu

show obvious concern about actual pronunciation (e.g., (d)). However,

taken together, they reveal a unanimous motive in their proposal:
while Saussure’s differential theory stipulates in principle that there is
nothing before differentiation and that all phonemes are generated
through phonetic differences, it does not offer any explanation as to
how phonemes are actually generated; consequently, all the above
criteria, including those that draw on actual pronunciation, are
proposed as part of a practical mechanism which guides us in the
derivation of specific phonemes from the differences observed.
Thus, the question that faced Chao and his contemporaries was
intrinsically the same question that Merleau-Ponty 62 asked at the
beginning of Indirect Language and the Voices of Silence, namely, «how
there could be a difference of meaning» between items that «do not
have any meaning at all»? Of course, for Merleau-Ponty, the question
was necessary because from the phenomenological point of view, one
had to probe into how the elements of a structure are actually
generated through the structure’s originary differentiation. For the
linguists, on the other hand, the question was apparently more about
sound than meaning, and it was necessitated by the practical need for
a methodology in the abstraction of phonemes. However, once the
question was asked, both Merleau-Ponty and the linguists had to face
the basic task of showing how differences finally lead to specific
structures with their sounds and concepts, or, to put it more vividly, of
«setting the Saussurean ‘differences’ to work». To fulfil the task, as we
have seen, Merleau-Ponty chose to draw on his own distinction
between the “speaking language” and the “spoken language” as well
as Saussure’s “speech”. But could the lively process of differentiation
and generation satisfy the need of the linguists? In Chao’s proposal,
we have already seen a tendency to reach out for something more
“concrete”, “reliable”, or “objective”. As he saw it, pure
differentiation, or, in his own words, the “purely logical definition”
provided by the differential theory, was not enough and could be
misleading. For example, it could lead us to the “unfavoured”

62 MERLEAU-PONTY 1964a, 39.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 119

conclusion that English [h] and [ŋ] are of one phoneme simply because
[h] never occurs at the end of a syllable and [ŋ] never occurs at the
beginning of a syllable. In such cases, Chao suggested that we rely on
actual pronunciation. This means that for him, there exists something
external to pure differentiation that guarantees the generation of pho-
nemes, and that this external source is not to be considered in any
other differentiation, but simply taken as a prior existence. Thus,
Chao’s reliance on actual pronunciation signals a tendency to go
around differentiation to a pre-given “thing”. If the tendency was only
a methodological one in Chao’s proposal, it soon grew into a
philosophical return to objectivity in the other linguists.
In the same year when Chao published The Non-uniqueness of
Phonemic Solutions to Phonetic Systems, Morris Swadesh published The
Phonemic Principle, another article that was to become a classic in
phonology. In the article, Swadesh proposed the idea of “phoneme
classes”, which he defined as phonemes with «common phonetic
characteristics, relatively analogous positional variants, and relatively
similar ranges of distribution».63 He laid a special emphasis on the
importance of phonetic properties in the understanding of the
phoneme classes:

It is important to distinguish between the phonetic

differentiae of phoneme classes and psychologically
separable synchronous features. Thus nasalization is the
phonetic differentia which in French distinguishes the
nasalized vowel phonemes from their non-nasalized
parallels […] psychologically separable synchronous
phonemes include tonemes in tone languages and tasemes
(phonemes of stress) in stress languages […].64

These statements are nothing but normal to most linguists. But

philosophically, they imply a step forward from Chao towards the

63 SWADESH 1934, 121.

64 SWADESH 1934, 121.

Metodo Vol. 4, n. 2 (2016)

120 Lei Zhu

objectification of phonemic differentiation. In Chao’s proposal, each

phoneme was considered for its own phonetic property, but in
Swadesh’s approach, phonetic properties of the phonemes are brought
together as the basic defining features according to which phonemes
can be categorised into classes. Again, it should be stressed that these
phonetic properties are not considered in any differentiation
themselves, but taken as the accepted natural qualities of sounds.
Thus, the so-called “phoneme classes” are not really relevant to any
originary differentiation, but established on the basis of a series of
objective properties. Such a treatment is only one step away from
attributing the differences between phonemes themselves to such
properties.
A very similar case to Swadesh’s approach appeared in Nikolai
Trubetzkoy’s Principles of Phonology (Grundzüge der Phonologie, 1939),
which is commonly known as the most important work of classical
phonemics. In the book, while focusing on phonemic contrasts
following Saussure, Trubetzkoy tried to disclose the different patterns
of such contrasts, and found “phonic properties” to be the basis on
which these patterns are established. According to his analysis, the six
phonemes in Figure 1, for instance, fall into three groups of bilateral
oppositions (p-b, t-d, k-ɡ) and two groups of trilateral oppositions (p-t-
k, b-d-ɡ):

p ↔ t ↔ k
↕ ↕ ↕
b ↔ d ↔ ɡ

Figure 1 Oppositional relations between the phonemes /p/, /b/, /t/, /d/, /k/, and /ɡ/

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 121

But more than that, Trubetzkoy, drawing on his phonetic knowledge,

found the bilateral oppositions here to be equally founded on the
existence/absence of voicing, and the trilateral oppositions on the
variation of the place of articulation between the bilabial, the dental
and the velar. As he declared in the beginning of Chapter IV in
Principles of Phonology, where the discussion of these phonic properties
was about to start:

Much more important for the general classification of

phonological oppositions is the fact that these oppositions
are phonic oppositions. […] The problem as to how phonic
properties are placed in opposition with each other, that is,
what types of opposition result, was discussed in Chapter
III. […]

In Chapter III we operated with purely logical concepts. We

must now combine these logical concepts with acoustical
and articulatory, that is, with phonetic, concepts. For no
other discipline except phonetics can teach us about
individual sound properties.65

Therefore, to say that the opposition “p-b” is a bilateral opposition,

for instance, is only to disclose the manner in which the opposition is
formed. For Trubetzkoy, what is more essential here is the fact that the
opposition is guaranteed by the voiced-voiceless contrast which exists
in the real world and can be acoustically and articulatorily explained.
Again, this sounds like common knowledge to any linguist now, but
philosophically, this is taking us further away from Saussure’s
differential theory towards total objectification. For in Saussure’s
theory, the foundation of structure is differentiation alone and nothing
else – «neither ideas nor sounds existing prior to the linguistic

65 TRUBETZKOY 1969, 91.

Metodo Vol. 4, n. 2 (2016)

122 Lei Zhu

system»,66 but in Trubetzkoy’s phonology, this foundation could no

longer stand alone, but had to be supported itself by something more
“realistic” in the physical sense. Although Trubetzkoy had a keen
interest in oppositional relations, for him, these relations meant only
the manifold forms in which the “real” physical existences presented
themselves. Thus, differentiation no longer had the originary status as
it did in Saussure’s differential theory, but was reduced to the manner
in which the more essential physical properties expressed themselves.
Of course, these physical properties were themselves prior to
differentiation.
So, if Chao tried to complement the Saussurean differences with
actual pronunciation, and Swadesh to categorise the results of these
differences using phonetic features, Trubetzkoy practically cancelled
the status of differentiation as the foundation for the generation of
phonemes, and substituted in its place a world of scientific facts and
objective entities. As he stated in a letter to his colleague and friend
Roman Jakobson in 1930:

The essence lies in the so to speak 'intrinsic content' of the

correlation. Apparently any (or might it not be 'any')
phonological correlation acquires in the linguistic
consciousness the form of a contraposition of a certain
mark to its absence (or of the maximum of a certain mark to
its minimum). Thus, one of the terms of the correlation
necessarily proves to be 'positive', 'active', and the other
becomes 'negative', 'passive'.67

The idea here is that oppositional relations of phonemes should be

defined not only by concrete properties, but by pairs of such
properties as the basic units or “intrinsic content” of those relations.
Thus, all phonemic oppositions would finally boil down to the binary
variation of certain physical elements – a situation found in the

66 SAUSSURE 1983, 118.

67 JAKOBSON & WAUGH 1979, 90.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 123

bilateral oppositions discussed above. This is an ingenious way to

transform the analysis of the manifold oppositional relations on the
basis of the assumed scientific entities underlying the relations. As
Trubetzkoy died before he could even complete the Principles of
Phonology, the transformation was left to Jakobson to carry out.
In 1952, Jakobson published the Preliminaries to Speech Analysis: The
Distinctive Features and their Correlates, which he coauthored with
Gunnar Fant and Morris Halle. In this monumental work, it was
proposed, for the first time, that phonemic differences can be reduced
to a matrix of “distinctive features” – the most fundamental units of
differentiation which cannot be further broken down and whose
number is very limited:

The distinctive features are the ultimate distinctive entities

of language since no one of them can be broken down into
smaller linguistic units. The distinctive features combined
into one simultaneous or, as Twaddell aptly suggests,
concurrent bundle form a 'phoneme'.68

Originally, Jakobson, Fant and Halle proposed twelve pairs of such

features, all of which had clear acoustic and articulatory correlates.
This feature system was then revised by many other scholars. The
most widely accepted version today, represented by Kenstowicz, 69 has
around 20 features which are grouped into several categories by their
phonetic contents. According to such a system, if a language has
/m/, /n/, /l/, /f/, /s/, /p/, /b/, /t/, /d/, /k/ and /ɡ/ as its
consonant phonemes, the consonant inventory could be reduced to a
matrix of positive-negative variation founded on just five distinctive
features, as shown below:

68 JAKOBSON, FANT & HALLE 1952, 3.

69 Kenstowicz 1994.

Metodo Vol. 4, n. 2 (2016)

124 Lei Zhu

m n l f s p b t d k ɡ
sonorant + + + -- -- -- -- -- -- -- --
continuant -- -- + + + -- -- -- -- -- --
voiced + + + -- -- -- + -- + -- +
anterior + + + + + + + + + -- --
coronal -- + + -- + -- -- + + -- --

Table 1 An 11-phoneme consonant system based on 5 distinctive features

For its brilliant way to present phonemic differentiation through the

interplay between fundamental physical units, the distinctive feature
analysis has long been accepted as a classical model of the structuralist
approach. Culler70 even praised the analysis for «re-establishing links
between language and thought» on the basis of the Saussurean
traidition, «but at the level of fundamental structuring operations».
However, with structural relations reduced to distinctive features and
phonemes to “concurrent bundles” of such features, this analysis
actually represents a philosophical reverse of Saussure’s differential
theory through a straightforward denial of Saussure’s principle that
«the language includes neither ideas nor sounds existing prior to the
linguistic system, but only conceptual and phonetic differences arising
out of that system».71 On the surface, the distinctive feature analysis is
still about structure and differentiation, but at the bottom, what
determines everything has been turned into a series of fundamental
atoms, or, in Jakobson’s words, “ultimate distinctive entities” that
refuse further division. Thus, in the “distinctive feature” version of
“structuralism”, structure or differentiation is no longer the originary
source of everything, but quite the opposite: everything is now the
result of particular combinations of atom- or quantum-like elements
that are ultimately objective and physical “things”. Philosophically,

70 CULLER 1976, 87.

71 SAUSSURE 1983, 118.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 125

this is certainly not structuralism – or at least Saussurean structuralism

– any more.
Indeed, the emphasis on structural differences started to decline in
linguistics after the 1950’s even as lip service. In 1959, Halle proposed
in The Sound Pattern of Russian that “phoneme” be abolished altogether
as a redundant and superficial notion that causes unnecessary
complexities in phonological analysis. In 1968, his masterpiece The
Sound Pattern of English, which he coauthored with Chomsky, was
published, marking the formal incorporation of the distinctive feature
analysis into Chomsky’s distinctly rationalist framework. Although
both the above works had “sound pattern” in their titles, their
understanding of sound patterns differed remarkably from that of
Saussure. With “phoneme” abandoned, these patterns were now the
direct results of their underlying phonetic “atoms” or “quanta”
through various combinations and changes. Their incorporation into
Chomsky’s rationalist generative grammar was easy and natural, for
structures – phonological or syntactic – were ultimately generated
from pre-given existences in this paradigm, where the originary
differentiation was completely dismissed.
Also incorporated into generative grammar was an approach to
meaning developed from the application of distinctive features to
semantic studies. Just as phonemes were reduced to distinctive
features, concepts were broken down into basic semantic units known
as “sememes”. For example, the concept “girl” could be analysed into
a combination of [+animate], [+human], [--male] and [--adult]. Such an
idea laid the basis for the so-called structural semantics as well as the
semantic analysis in generative grammar and generative semantics.
Philosophically, this could be regarded as a signal of the complete
retreat of Saussurean structuralism from linguistics, as the structure of
the sign, with the dominance of the idea, was now broken down into
smaller self-dependent entities not only at the sound level, but at the
conceptual level as well.
The application of distinctive features to semantics takes us back to
the comparison between Merleau-Ponty and the linguists, for its

Metodo Vol. 4, n. 2 (2016)

126 Lei Zhu

concern with conceptual structures has put them in more direct

contrast. But as we have mentioned, even in phonemic analysis, the
post-Saussurean linguists shared intrinsically the same question as the
one confronting Merleau-Ponty, namely, if there are only differences in
the beginning and nothing else, how are the elements of the structure
derived from the originary differentiation in the first place? While
Merleau-Ponty was mainly focused on the derivation of concepts, the
linguists started from the derivation of phonemes before they moved
on to concepts. But no matter dealing with phonemes or concepts, as
we have seen, the linguists’ approach was sharply different from that
of Merleau-Ponty, who tried to make sure not to fall into objectivity or
subjectivity: the linguists tried to look beneath structural
differentiation for more “reliable” guarantees, until they found them
in what they believed to be the primordial and indivisible units of
pronunciation and meaning.
Thus, while Saussure’s differential theory was turned by Merleau-
Ponty into a part of his phenomenology, in linguistics it underwent a
transition to what Merleau-Ponty would call “traditional philosophy”.
As a last note to our discussion of the latter transition, it may be
worthwhile to point out that the generative phonologist Sanford
Schane argued in an influential article titled The Phoneme Revisited that
the notion of “phoneme” should not be abandoned, because it could
still be useful in the explication of certain phenomena arising out of
the influence of superficial phonetic contrasts on the speaker. But at
the same time, to keep his theory consistent, Schane suggested that
“phoneme” be regarded as an epiphenomenon in phonology.
Believing that he had found the rightful place for “phoneme” in a
world where contrasts were no longer fundamental, he declared:

The phoneme was the offspring of structuralism, the pride

and joy of post-Bloomfieldian linguistics. Since then the
child has been abandoned. Yet some of us may have felt
guilty about disinheriting the child. As generativists, if we
acknowledged him, then it was as an illegitimate child.

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 127

Perhaps we can now recognize the little bastard for what he

really is.72

The same, actually, could be said of structure and differentiation in

general.

5. Conclusion
In this paper, I traced the development of Saussure’s differential
theory of the linguistic sign from its formation in Saussure’s early
studies to its maturation in the Course in General Linguistics. Then, I
examined the transformation of the theory in Merleau-Ponty’s
philosophy as well as its interpretation in the works of the post-
Saussurean linguists.
As I have shown, Saussure’s understanding of the fundamental role
of differences was deeply rooted in his historical linguistic research, in
which he made it a principle to put sound patterns before actual
pronunciations. His elaboration on the differential theory in the Course
in General Linguistics was in a sense an extension of the principle to
concepts. The statement in the Course in General Linguistics that no
sound or concept exists prior to the linguistic system, in which there
are only phonetic and conceptual differences, inspired both Merleau-
Ponty and the post-Saussurean linguists. Merleau-Ponty, from his own
phenomenological perspective, found Saussure’s differential theory a
useful tool in achieving his reflective goal without incurring a
subjective tendency. The linguists, on the other hand, made the theory
a fundamental principle in the newly-developed area of phonemics,
which set the model for the other branches of linguistics for a long
time.
However, Saussure did not consider or explain how specific
phonemes and concepts could actually derive from the originary
differentiation, or, in other words, how the elements of a structure

72 SCHANE 1971, 520.

Metodo Vol. 4, n. 2 (2016)

128 Lei Zhu

could arise if there are only differences in the beginning, and this
caused a problem for both Merleau-Ponty and the linguists. Merleau-
Ponty, while trying to turn the differential theory into a new ontology
and the foundation of a new philosophy of history, felt it necessary to
account for the generation of concepts from differentiation so as to
prevent structure or structural differences from being objectified into
static beings. What he did was to turn to his own distinction between
the “speaking language” and the “spoken language” as well as
Saussure’s “speech”, and too highly stressed the act of speaking as the
locus where concepts are generated through differentiation. In
contrast, the linguists found the question about the generation of
structural elements compelling because without answering it, they
would not be able to establish a methodology for the actual abstraction
of phonemes from raw data. What they chose to do, as scientists, was
to reach out for more “concrete” and “reliable” sources, which they
finally found in some basic phonetic properties or “distinctive
features”, understood as physical existences. Consequently, they
reduced linguistic structures to their underlying “atoms” that existed
before the structures themselves. This meant, however, that structural
differentiation actually lost its originary status in linguistics and gave
way to objective and physical “things”.

References
BARTHES, R. 1968. Elements of Semiology. Trans. Lavers, A. & Smith, C.
New York: Hill and Wang.
CARMAN, T. & HANSEN, M. 2004. «Introduction». In Carman, T. &
Hansen, M. (eds.) The Cambridge Companion to Merleau-Ponty.
Cambridge: Cambridge University Press, 1-25.
CHAO, Y.-R. [1934] 2006. «The Non-uniqueness of Phonemic Solutions

Metodo Vol. 4, n. 2 (2016)

Interpreting Saussure's Differential Theor 129

of Phonetic Systems». In Wu, Z.-J. & Zhao, X.-N. (eds.) Linguistic

Essays by Yuenren Chao. Beijing: The Commercial Press, 221-79.
CLACKSON, J. 2007. Indo-European Linguistics: An Introduction.
Cambridge: Cambridge University Press.
CULLER, J. 1975. Structuralist Poetics. London & New York: Routledge.
— 1976. Saussure. London: Fontana/Collins.
DOSSE, F. 1997. History of Structuralism, Volume I: The Rising Sign,
1945-1966. Trans. D. Glassman. Minneapolis: University of
Minnesota Press.
GUTTING, G. 2001. French Philosophy in the Twentieth Century.
Cambridge: Cambridge University Press.
HASS L. 2008. Merleau-Ponty’s Philosophy. Bloomington &
Indianapolis: Indiana University Press.
JAKOBSON, R., FANT, G. & HALLE, M. 1952. Preliminaries to Speech
Analysis: The Distinctive Features and Their Correlates.
Cambridge: The MIT Press.
JAKOBSON, R. & WAUGH, L. 1979. The Sound Shape of Language.
Bloomington: Indiana University Press.
JOSEPH, J. 2012. Saussure. Oxford: Oxford University Press.
KAGANOI, S. 1998. «Merleau-Ponty and Saussure: On the Turning Point
of Merleau-Ponty's Thinking». In Tymieniecka, A.-T. & Matsuba, S.
(eds.) Immersing in the Concrete: Maurice Merleau-Ponty in the
Japanese Perspective. Dordrecht: Springer, 151-72.
KENSTOWICZ, M. 1994. Phonology in Generative Grammar. Oxford:
Blackwell.
KURYŁOWICZ, J. 1927. «ə indoeuropéen et ḫ Hittite». In Taszycki, W. et
al. (eds.) Symbolae Grammaticae in Honorem Ioannis
Rozwadowski, Vol. 1. Cracow: Drukarnia Uniwersytetu
Jagiellonskiego, 95-104.
LAGEUX, M. 1965. «Merleau-Ponty et la linguistique de Saussure».
Dialogue 4.3, 351-64.
MALDINEY, H. 2000. «Flesh and Verb in the Philosophy of Merleau-
Ponty». In Evans, F. & Lawlor, L. (eds.) Chiasms: Merleau-Ponty’s
Notion of Flesh. Albany: State University of New York Press, 51-76.

Metodo Vol. 4, n. 2 (2016)

130 Lei Zhu

MERLEAU-PONTY, M. 1962. Phenomenology of Perception. Trans. Smith,

C. London & New York: Routledge.
— 1963. In Praise of Philosophy. Trans. Wild, J. & Edie, J. Evanston:
Northwestern University Press.
— 1964a. Signs. Trans. McCleary, R. Evanston: Northwestern
University Press.
— 1964b. The Primacy of Perception. Ed. Edie, J. Evanston:
Northwestern University Press.
— 1968. The Visible and the Invisible. Trans. Linguis, A. Evanston:
Northwestern University Press.
— 1973. The Prose of the World. Trans. O'Neill, J. Evanston:
Northwestern University Press.
MUGDAN, J. 2011. «On the Origin of the Term Phoneme».
Historiographia Linguistica 38.1/2, 85-110.
RICOEUR, P. 1967. «New Developments in Phenomenology in France:
The Phenomenology of Language». Social Research 34.1, 1-30.
ROBINS, R. 1997. A Short History of Linguistics. London & New York:
Longman.
SAUSSURE, F. de. [1878] 1922. «Mémoire sur le système primitif des
voyelles dans les langues indo-européennes». In Recueil des
publications scientifiques de Ferdinand de Saussure. Lausanne:
Librairie Payot, 1-268.
— 1983. Course in General Linguistics. Trans. Harris, R. London:
Duckworth.
SCHMIDT J. 1985. Maurice Merleau-Ponty: Between Phenomenology
and Structuralism. New York: St. Martin’s Press.
SWADESH, M. 1934. «The Phonemic Principle». Language 10.2, 117-29.
TRUBETZKOY, N. 1969. Principles of Phonology. Trans. Baltaxe C.
Berkeley & Los Angeles: University of California Press.

Metodo Vol. 4, n. 2 (2016)