markedness (n.)

An analytic principle in LINGUISTICS whereby pairs of linguistic FEATURES, seen as OPPOSITIONS, are given different values of POSITIVE (marked) and NEUTRAL or NEGATIVE (unmarked). In its most general sense, this distinction refers to the presence versus the absence of a particular linguistic feature. There is a formal feature marking plural in most English nouns, for example; the plural is therefore ‘marked’, and the singular is ‘unmarked’. The reason for postulating such a relationship becomes clear when one considers the alternative, which would be to say that the opposed features simply operate in parallel, lacking any directionality. Intuitively, however, one prefers an analysis whereby dogs is derived from dog rather than the other way round – in other words, to say that ‘dogs is the plural of dog’, rather than ‘dog is the singular of dogs’. Most of the theoretical discussion of markedness, then, centres on the question of how far there is intuitive justification for applying this notion to other areas of language (cf. prince/princess, happy/unhappy, walk/walked, etc.).

One of the earliest uses of the notion was in PRAGUE SCHOOL PHONOLOGY, where a sound would be said to be marked if it possessed a certain DISTINCTIVE FEATURE (e.g. VOICE), and unmarked it if lacked it (this unmarked member being the one which would be used in cases of NEUTRALIZATION). In GENERATIVE phonology, the notion developed into a central criterion for formalizing the relative NATURALNESS of alternative solutions to phonological problems. Here, evidence from frequency of occurrence, HISTORICAL LINGUISTICS and language ACQUISITION is used to support the view that marking is a basic principle for assigning UNIVERSAL (and possibly innate) values to PHONETIC features (by contrast with the language-specific, phonological approach of the Prague School). The distinctive features are each assigned marking values, e.g. [+voice] is seen as marked, [−voice] as unmarked. SEGMENTS, in this view, can then be seen as combinations of marked or unmarked features, and thus be compared with each other, e.g. /a/ is the maximally unmarked vowel because it is [−high], [−back] and [−round];  is more complex because it is [+low] and [+round], and so on. In later phonological theory (e.g. in UNDERSPECIFICATION theory), the notion of markedness took on a critical status. Based on the view that the unmarked value of a feature is the normal, neutral state of the relevant articulator, some approaches assert that only one value need be present in the UNDERLYING REPRESENTATION; the other can be predicted by a CONTEXT-free RULE which mirrors the relevant markedness statement. For example, [ ] ⇒ [−nasal] would represent the notion that segments are normally oral. The rule would insert [−nasal] by default only in segments lacking a nasal value. Such rules are known as ‘markedness-based context-free redundancy rules’.

Several other interpretations of the notion of marking are found in the literature, where the concept of ‘presence v. absence’ does not readily apply. One interpretation relates marking to frequency of occurrence, as when one might say a FALLING INTONATION pattern was unmarked, compared with a rising one, because it is more common. Another is found in the SEMANTIC analysis of LEXICAL ITEMS, where pairs of items are seen as unmarked and marked respectively, on the grounds that one member is more specific than the other (e.g. dog/bitch, where the latter is marked for sex – one can say male/female dog, but these ADJECTIVES are inapplicable with bitch). A third, related sense occurs when the DISTRIBUTION of one member of an opposition is restricted, compared with the other: the restricted item is then said to be marked – several COMPARATIVE SENTENCES illustrate this, e.g. How tall is John? (where How short is John? is abnormal). In later GENERATIVE linguistics, a more general theory of markedness emerged. Here, an unmarked property is one which accords with the general tendencies found in all languages; a marked property is one which goes against these general tendencies – in other words, it is exceptional (a RELATIVE UNIVERSAL). Markedness in this sense can be represented as a continuum along which language-universal and language-specific properties can be related. A highly unmarked property is one which makes a strong claim to universal status; a highly marked property is one which makes a weak universal claim. A universal which is strongly represented in a particular language makes that language highly unmarked in that respect, and vice versa. For example, in relation to the proposed phonological universal that words must start with a consonant+ vowel structure (CV), some languages (e.g. Yawelmani) totally satisfy this universal, whereas others (e.g. English) do not; English is therefore more marked than Yawelmani, in this respect. In OPTIMALITY THEORY, the RANKING of CONSTRAINTS and constraint VIOLATIONS allows the notion of markedness to be encoded directly into the model.