Ferdinand de Saussure

Let's start by talking about structuralism in general as a philosophical stance or worldview. Structuralists are interested in the interrelationship between UNITS, also called "surface phenomena," and RULES, which are the ways that units can be put together. An example is Tinkertoys. The "units" in a tinkertoy set are all the parts in the box: the various colored rods of different lengths, the various kinds of connectors and wheels and attachments; the "rules" of tinkertoy construction is that rods go into holes. That's the structure of tinkertoys: everything you can make out of tinkertoys, whatever that may be, is made by using the units according to the rules. A structuralist analysis of tinkertoys wouldn't look at what you made (a building, a race car, a windmill, etc.) but would look only at the structure governing every possible combination of tinkertoy elements. And that structure is that rods go into holes.

That's what structuralist analysis does, whatever it's analyzing: looks at the units of a system, and the rules that make that system work, without regard for any specific content. In language, for instance, structuralists (like Saussure) the units are words (or, actually, the 31 phonemes which make all the sounds of words in English) and the rules are the forms of grammar which order words. In different languages the grammar rules are different, as are the words, but the structure is still the same in all languages: words are put together within a grammatical system to make meaning.

An example of this idea of structure can be found in the game of "Mad Libs." In class I read an example, which asked for various nouns, adjectives, verbs, proper names, and exclamations. When plugged into a story, these randomly chosen parts of speech made a very silly narrative-- but one which was recognizable as a narrative because the parts of speech were appropriately placed: nouns went where nouns go, and verbs where verbs go, etc. In a sentence, any noun can replace any other noun and not change the grammatical structure: the sentence "My pencil ate my PT Cruiser" might not make any rational sense, but it's recognizable as a sentence because the parts of speech are all in the right places. Here's an example of this using literature. I'll give you three characters: princess, stepmother, and prince. Now you tell me the story. Many of you said "Cinderella," and others came up with other story titles. From a structuralist point of view, Cinderella is the same story as Snow White and as lots of other Disney stories and fairy tales: a princess is persecuted by a stepmother and rescued (and married) by a prince. The "units" here are the characters, and the "rules" are: stepmothers are evil, princesses are victims, and princes and princesses have to marry. Whatever details or added elements you supply, the basic structure of this story is always the same. And that's exactly what structuralist analyses of literature (or myth or other forms of narrative) are analyzing.

Structuralists believe that the underlying structures which organize units and rules into meaningful systems are generated by the human mind itself, and not by sense perception. As such, the mind is itself a structuring mechanism which looks through units and files them according to rules. This is important, because it means that, for structuralists, the order that we perceive in the world is not inherent in the world, but is a product of our minds. It's not that there is no "reality out there," beyond human perception, but rather that there is too much "reality" (too many units of too many kinds) to be perceived coherently without some kind of "grammar" or system to organize and limit them.

So structuralism sees itself as a science of humankind, and works to uncover all the structures that underlie all the things that humans do, think, perceive, and feel--in mathematics, biology, linguistics, religion, psychology, and literature, to name just a few disciplines that use structuralist analyses.

Structuralist analysis posits these systems as universal: every human mind in every culture at every point in history has used some sort of structuring principle to organize and understand cultural phenomena. For instance, every human culture has some sort of langauge, which has the basic structure of all language: words/phonemes are combined according to a grammar of rules to produce meaning. Every human culture similarly has some sort of social organization (like a kind of government), some sort of system for who can marry whom (usually referred to as a kinship system), and some sort of system for exchanging goods (usually referred to as an economic system). All of these organizations are governed, according to structuralist analyses, by structures which are universal.

For a more formal definition: a structure is any conceptual system that has the following three properties:

  • Wholeness. This means that the system functions as a whole, not just as a collection of independent parts. With the tinkertoy set, it's hard to play with just the individual items; you need the whole set, with all the rods and holes, and the rules they follow, in order to make stuff at all.
  • Transformation. This means that the system is not static, but capable of change. New units can enter the system, but when they do they're governed by the rules of the system. With the tinkertoys, I can substitute a blue wheel for a yellow one, or an orange rod for a purple one, but rods still go into holes in order to create something. With tinkertoys, you could add any rods (of the right diameter) and any holes (ditto) and the system still works. Another example is the word "office"--normally it's a noun, but a Kinko's commercial has made it a verb, as in "a new way to office." The commercial creates a new word, "to office," and we know what it means because the structure it fits into hasn't changed.
  • Self-Regulation. This is related to the idea of transformation. You can add elements to the system, but you can't change the basic structure of the system no matter what you add to it. The transformations of a system never lead to anything outside the system. We can add things to the tinkertoy set and never alter the fundamental rule that rods go into holes. (Stay tuned, though; poststructuralist theories will challenge this point).

  • Now then. On to Saussure, who is provides us with a structuralist analysis of language as a signifying system.

    Why are we studying Saussure, a linguistic theorist, in a literature class? When we discard the assumptions of liberal humanism, we start our new conceptions of how literature operates by noting that, first and foremost, literature is made of language; to understand how literature works, we must therefore have some ideas about how language itself works. Saussure, as a structuralist, is interested in language as a system or structure. His ideas apply to any language--English, French, Farsi, computer languages--and to anything we can call a "signifying system" (more on what this is later). He describes the structures within any language which make meaning possible, but he's not interested in what particular meanings get created. Like all structuralists, he's not interested in the details of what fills up the structure, the specifics of speech or writing, but only in the design of the structure itself.


    Language is based on a NAMING process, by which things get associated with a word or name. Saussure says this is a pretty naive or elementary view of language, but a useful one, because it gets across the idea that the basic linguistic unit has two parts.

    Those two parts Saussure names the "concept" and the "sound image". The sound image is not the physical sound (what your mouth makes and your ear hears) but rather the psychological imprint of the sound, the impression it makes. An illustration of this is talking to yourself--you don't make a sound, but you have an impression of what you're saying.

    The linguistic SIGN (a key word) is made of the union of a concept and a sound image. The union is a close one, as one part will instantly conjure the other; Saussure's example is the concept "tree" and the various words for tree in different languages. When you are a speaker of a certain language, the sound image for tree in that language will automatically conjure up the concept "tree."  The MEANING of any SIGN is found in the association created between the sound image and the concept: hence the sounds "tree" in English mean the thing "tree." Meanings can (and do) vary widely, but only those meanings which are agreed upon and sanctioned within a particular language will appear to name reality. (More on this as we go on).

    A more common way to define a linguistic SIGN is that a SIGN is the combination of a SIGNIFIER and a SIGNIFIED. Saussure says the sound image is the SIGNIFIER and the concept the SIGNIFIED. You can also think of a word as a signifier and the thing it represents as a signified (though technically these are called sign and referent, respectively).

    The SIGN, as union of a SIGNIFIER and a SIGNIFIED, has two main characteristics.

    1. The bond between the SIGNIFIER (SFR) and SIGNIFIED (SFD) is ARBITRARY. There is nothing in either the thing or the word that makes the two go together, no natural, intrinsic, or logical relation between a particular sound image and a concept. An example of this is the fact that there are different words, in different languages, for the same thing. Dog is "dog" in English, "perro" in Spanish, "chien" in French, "Hund" in German.

    This principle dominates all ideas about the STRUCTURE of language. It makes it possible to separate the signifier and signified, or to change the relation between them. (This makes possible the idea of a single signifier which could be associated with more than one signified, or vice-versa, which makes AMBIGUITY and MULTIPLICITY OF MEANING possible.)

    Language is only one type of semiological system (the word "semiological," like the word "semiotic," comes from the Greek word for "sign"). Any system of signs, made up of signifiers and signifieds, is a semiotic or SIGNIFYING SYSTEM. Think, for example, of football referee signals, baseball signs, astrological signs. Any time you make up a secret code or set of signals you are making your own signifying system.

    There may be some kinds of signs that seem less arbitrary than others. Pantomime, sign language, gestures (what are often called "natural signs") seem to have a logical relation to what they represent. The tomahawk chop used by Atlanta Braves fans, for example, seems to imitate the action of chopping, and thus would be the most "natural" way to designate the idea of chopping. But Saussure insists that ALL SIGNS ARE ARBITRARY; the tomahawk chop only has meaning because a community has agreed upon what the gesture signifies, not because it has some intrinsic meaning.

    Saussure discusses whether symbols, such as the use of scales for the idea of justice, are innate or arbitrary, and decides that these too are arbitrary, or based on community agreement. He also dismisses onomatopoeia (words that sound like what they mean, like "pop" or "buzz") as still conventional, agreed-upon approximations of certain sounds. Think, for example, about the sounds attributed to animals. While all roosters crow pretty much the same way, that sound is transcribed in English as "cock-a-doodle-do" and in Spanish as "cocorico." Interjections also differ. In English one says "ouch!" when one bangs one's finger with a hammer; in French one says "Aie!" (Curse words work the same way. Come up with your own examples).

    Admittedly, Saussure is not very interested in how communities agree on fixing or changing the relationships between signifiers and signifieds. Like all structuralists, he focuses on a SYNCHRONIC analysis of language as a system or structure, meaning that he examines it only in the present moment, without regard to what its past history is, or what its future may be. (Analyses which do take time into account, and look at the history of changes within a structure, are called DIACHRONIC).

    2. The second characteristic of the SIGN is that the signifier (here, meaning the spoken word or auditory signifier) exists in TIME, and that time can be measured as LINEAR. You can't say two words at one time; you have to say one and then the next, in a linear fashion. (The same is true for written language: you have to write one word at a time (though you can write over an already written word) and you generally write the words in a straight line).

    This idea is important because it shows that language (spoken language, anyway) operates as a linear sequence, and that all the elements of a particular sequence form a chain. The easiest example of this is a sentence, where the words come one at a time and in a line, one after the other, and because of that they are all connected to each other.


    According to Saussure's picture (p. 649a), thought is a shapeless mass, which is only ordered by language. One of the questions philosophers have puzzled over for centuries is whether ideas can exist at all without language. (Think, for example, about Helen Keller before she learned language--did she think?) Saussure says no ideas preexist language; language itself gives shape to ideas and makes them expressible. In other words, from Saussure's point of view, thought cannot exist without language. (This leads to an important structuralist and post-structuralist idea, which is that language shapes all our conceptions of ourselves and our reality. More on this later). Sound is no more fixed than thought, though sounds can be distinguished from each other, and hence associated with ideas. Sounds then serve as signifiers for the ideas which are their signifieds. Signs, in this view, are both material/physical (like sound) and intellectual (like ideas). This is important to Saussure because he wants to insist that language is not a thing, a substance, but a form, a structure, a system. His image is that thought and sound are like the front and back of a piece of paper (and the paper is the linguistic sign); you can distinguish between the two, but you can't separate them. Saussure (and other structuralist and post-structuralist theorists) talk about the system of language as a whole as LANGUE (from the French word for language), and any individual unit within that system (such as a word) as a PAROLE. Structuralist linguistics is more interested in the LANGUE than in any PAROLE. (Peter Barry, in Beginning Theory, talks about literary systems, like genre categories, as a form of LANGUE, and individual literary texts as examples of PAROLE). The arbitrary nature of the sign explains why language as a system (LANGUE) can only arise in social relations. It takes a community to set up the relations between any particular sound image and any particular concept (to form specific PAROLES). An individual can't fix VALUE for any signifier/signified combination. You could make up your own private language, but no one else would understand it; to communicate, two or more people have to agree on what signifiers go with what signifieds. (And again, Saussure as a structuralist is not really interested in how this happens. Other theorists of language, such as 18th century philosopher Jean-Jacques Rousseau, focus on how these agreements come about).

    VALUE is thus defined as the collective meaning assigned to signs, to the connections between sfrs and sfds.

    The VALUE of a sign is determined, however, not by what signifiers get linked to what particular signifieds, but rather by the whole system of signs used within a community. VALUE is the product of a system or structure (LANGUE), not the result of individual sfr-sfd relations (PAROLE).

    Saussure distinguishes between VALUE and SIGNIFICATION. SIGNIFICATION is what we commonly think of as "meaning," the relationship established between a signifier and a signified. VALUE, by contrast, is the relation between various SIGNS within the signifying system. As Saussure says on p. 650b: "Language is a system of interdependent terms in which the value of each term results solely from the simultaneous presence of the others."

    VALUE is always composed of two kinds of comparisons among elements in a system. The first is that dissimilar things can be compared and exchanged, and the second is that similar things can be compared and exchanged. A good example of this is money. A dime is a signifier connected to a signified of 10 cents of something. The VALUE of a dime is established because it can be exchanged for something dissimilar--a piece of gum--or something similar--ten pennies. (Coins are also good examples of the arbitrary nature of signs. A dime is worth 10 cents because we all agree that it is, not because the materials in the coin have some absolute value of 10 cents).

    Words work the same way. A word can be "exchanged" for something similar--another word, a synonym--or for something dissimilar--an idea, for example. In both cases (coin or word), it is the system itself which creates value, and sets up the ways that exchanges can be made. A signifier, such as a coin or a word, when considered alone, has only a limited relation to its own signified; when considered as part of a system, a signifier has multiple relations to other signifiers in the system.

    The most important relation between signifiers in a system, the relation that creates VALUE, is the idea of DIFFERENCE. One signifier has meaning within a system, not because it's connected to a particular signified, but because it is NOT any of the other signifiers in the system. The word "cat" has meaning, not because of the animal it's associated with, but because that word is not "hat" or "bat" or "car" or "cut."

    You might think about the letters of the alphabet in this context. The sound "t-t-t-t", made with the tip of the tongue against the teeth, is represented in English with the symbol "T." Because the connection between sound and concept, or signifier and signified, is ARBITRARY, that sound "t-t-t-t" could just as easily be represented by another symbol, such as "D" or "%". Further, within the alphabet, "T" has meaning because it is NOT "A" or "B" or "X." Saussure calls this a negative value, wherein something has meaning or value because it is NOT something else within a system. (Positive value, on the other hand, is established in the sfr/sfd connection; a sign has positive value in and of itself because of the connection of its two parts, but has negative value within a signifying system). Another good example of this is the digital languages recognized by computers, which consist of two switch positions, off and on, or O and 1. O has meaning because it is not 1, and 1 has meaning because it is not 0.

    The system of linguistic units depends thus on the idea of DIFFERENCE; one unit has VALUE within the system because it is not some other unit within the system. As the computer example shows, this idea of DIFFERENCE depends upon the idea of BINARY OPPOSITES. To find out what a word or sign is not, you compare it to some other word or sign. (And because language exists in time and space, you can only do this comparison one word at a time, hence always forming binary pairs, pairs of two.) A binary pair shows the idea of difference as what gives any word value: in the pair cat/cats, the difference is the "s"; what makes each word distinct is its difference from the other word. (Saussure uses the example of Nacht and N[[questiondown]][[questiondown]]chte, p. 653).


    In this section, Saussure says more about how he thinks the structure of language, or of any signifying system, operates. Everything in the system is based on the RELATIONS that can occur between the units in the system. These relations, as we've already noted, consist mainly of relations of DIFFERENCE. In this section Saussure talks more about the rules that may connect units together.

    The most important kind of relation between units in a signifying system, according to Saussure, is a SYNTAGMATIC relation. This means, basically, a LINEAR relation. In spoken or written language, words come out one by one (see above, the second characteristic of the linguistic sign). Because language is linear, it forms a chain, by which one unit is linked to the next.

    An example of this is the fact that, in English, word order governs meaning. "The cat sat on the mat" means something different than "The mat sat on the cat" because word order--the position of a word in a chain of signification--contributes to meaning. (The sentences also differ in meaning because "mat" and "cat" are not the same words within the system).

    English word order has a particular structure: subject-verb-object. Think of this sentence: "The adjectival noun verbed the direct object adverbially." Other languages have other structures; in German, that sentence might be "The adjective noun auxiliary verbed the direct object adverbially main verb." In French it might be "The noun adjective verbed adverbially the direct object ." In Latin, word order doesn't matter, since the meaning of the word is determined, not by its place in the sentence, but by its cases (nominative, ablative, etc.)

    Combinations or relations formed by position within a chain (like where a word is in a sentence) are called SYNTAGMS. Examples of SYNTAGMS can be any phrase or sentence that makes a linear relation between two or more units: under-achiever; by the way; lend me your ears; when in the course of human events.

    The terms within a syntagm acquire VALUE only because they stand in opposition to everything before or after them. Each term IS something because it is NOT something else in the sequence. Again, think of coins: a dime is a dime because it's not a quarter or a nickel or a penny or a $100 bill.

    SYNTAGMATIC relations are most crucial in written and spoken language, in DISCOURSE, where the ideas of time, linearity, and syntactical meaning are important. There are other kinds of relations that exist outside of discourse.

    Signs are stored in your memory, for example, not in syntagmatic links or sentences, but in ASSOCIATIVE groups. The word "education", for example, may get linked, not to verbs and adjectives, but to other words that end in "-tion":education, relation, association, deification. You may store the word education" with other words that have similar associations: education, teacher, textbook, college, expensive. Or you may store words in what looks like a completely random set of linkages: education, baseball, computer games, psychoanalysis (things I like). The idea of ASSOCIATIVE groups or linkages makes me think of pigeonholes, and what pigeonholes I put certain words or ideas in; when I pull out that word or idea, all the other things in that pigeonhole come tumbling out with it.

    ASSOCIATIVE relations are only in your head, not in the structure of language itself, whereas SYNTAGMATIC relations are a product of linguistic structure.

    Think of the columns of a building (or the rods in a Tinker-Toy "building"). The columns form syntagmatic, or structural, relation when you think about where in the building the columns are, what they support, what they're connected to. The columns form associative relations when you think of what else the columns make you think of: phallic symbols, rockets, popsicles, or whatever.

    syntagmatic relations are important because they allow for new words--neologisms--to arise and be recognized and accepted into a linguistic community. "To office," for example (now used in a Kinko's commercial) has meaning because the noun "office" can be moved to the position of verb, and take on a new syntagmatic position and relation to other words. Associative relations are important because they break patterns established in strictly grammatical/linear (syntagmatic) relations and allow for metaphoric expressions.

    Some of the ideas about structuralism in this lecture were inspired by Lois Tyson, Critical Theory Today: A User-Friendly Guide (New York and London: Garland Publishing Inc., 1999) pp. 197-200.
    All page references in this lecture refer to Ferdinand de Saussure's "Course in General Linguistics" in Adams and Searle, ed., Critical Theory Since 1965.
    Suggestions for Further Reading:
  • Richard and Fernande de George, eds., The Structuralists from Marx to Levi-Strauss, 1972.

  • John Sturrock, Structuralism, 1986.

  • Roy Harris, Reading Saussure, 1987

  • Terrence Gordon, Saussure for Beginners, 1996.
  • All materials on this site are written by, and remain the property of, Dr. Mary Klages, Associate Professor of English, University of Colorado at Boulder. You are welcome to quote from this essay, or to link this page to your own site, with proper citation and attribution: http://www.colorado.edu/English/ENGL2012Klages/saussure.html