The true nature of “markedness” in the history of phonology is highly uncertain, in that the term is used to refer to a wide array of facts about language, and there is little agreement over what the term even refers to, much less whether it is a valid concept. This paper reviews certain applications of that concept in phonology, in search of some unity behind “markedness”. I show that “markedness” is about two unrelated things: formal properties of language, and functional probability of occurrence. Much effort has been put into forcing these two conceptions under a single computation umbrella, and that effort bears significant responsibility for the development of substance-dependent theories of grammar. As for whether “markedness” is a worthy topic of investigation, it is argued that the original formal question underlying markedness is still worth scrutiny in the theory of grammar: what is the nature of phonological features?
Halle’s (1959) argument against a distinction between morphophonemic and phonemic rules can be understood as an argument against the relevance of contrast to phonology. After adducing further arguments against a role for contrast, the paper provides a simple contrast-free analysis of the classic problem of the voicing behavior of Russian /v/. This segment undergoes voicing assimilation (like other obstruents), but does not trigger it (thus acting like the sonorants). In contrast to a long history of treating /v/ as a covert sonorant, the paper attributes the behavior of Russian /v/, which surfaces always as an obstruent, to underspecification with respect to the feature Voice.
The 21st century has brought major advances in our understanding of sound patterns, their phonetic basis, and their cross-linguistic diversity. Properties that were once thought to be universal, from phonological features to prosodic units like the syllable, have shown themselves to be highly variable across languages, and to sometimes fail at the very specific organizational role they are meant to play within sound systems. The goal of this chapter is to summarize evidence from the diversity of phonological systems that may inform areas of disagreement in modeling phonological grammars, with special attention to the locus of explanation in phonological theory. General arguments against phonological universals and phonological markedness in grammar are presented, including distinctive features, the sonority scale, and the prosodic hierarchy. Arguments for language-specific sound patterns and extra-grammatical explanations are also presented, including rare phoneme inventories, rare phonotactics, and distinct modality-specific properties of spoken versus signed languages.
Empty categories – positions in phonological representations that have no direct phonetic counterpart – are (still) controversial in phonological theory. In this paper we give the main arguments for assuming such positions and we furthermore establish a markedness hierarchy for empty positions: some of them are stronger (‘more marked’) than others, and we can derive this from a combination of Element Theory and Turbidity Theory. We illustrate our point with Italian and Dutch dialects, and point out that the phonological hierarchy of empty positions may correspond to a hierarchy of syntactic positions.
In terms of the roles of heads and dependents, there is a mismatch between phonology and syntax. In phonology, heads are important in both structural and informational terms. In syntax, on the other hand, heads have a different function: syntactic heads, like phonological heads, have an important structural role because they license dependent structure, but unlike phonological heads their informational role is relatively unimportant because they usually bear less linguistic (e.g. lexical) information than dependents. In order to achieve a greater degree of uniformity between phonology and syntax, this chapter proposes a reassessment of the roles of heads and dependents in phonology. Contrary to the widespread view, it is argued that heads in phonology, like those in syntax, are structurally important but lexically unimportant whereas dependents are structurally unimportant but lexically important. This view is supported not only by segmental distribution patterns but also by the size of the modulated carrier signal (rather than by the more standard phonetic measure, the sonority scale).
As a linguistic concept, markedness is in trouble (Hale & Reiss 2000, 2008; Blevins 2004; Haspelmath 2006; Samuels 2011; Reiss 2017). Accepting many of these criticisms and conclusions, this chapter, nevertheless, claims that there is space for markedness in formal phonology. By examining syllable structure from a highly representational viewpoint, I conclude that ‘markedness by complexity’ does have explanatory value and that markedness is explicable in terms of ‘structural complexity’ and ‘length of description’. The core demonstration is based on Charette’s (1990, 1991, 1992) formal typology of consonant clusters (CCs). These papers contain important implicational universals that Charette relates directly to her representations. My contribution will be to enrich the typology and reorganize those principles and parameters into a decision tree or parameter hierarchy that derives the implicational universals. Moreover, the number of parameter settings (the depth in the parameter hierarchy) increases the markedness of the resultant grammar. Crucially, each parametric ‘yes’ setting corresponds to an extra empty phonological category or extra ability to license, or reflects the distance between heads and dependents in the representation. For this reason, markedness is not merely a ‘metaphor’ ‘for a cognitive state’ (Haspelmath 2006), it is directly convertible into linguistic categories. This markedness is not part of the online computation of forms (contra Optimality Theory) and markedness statements cannot be re-ranked to obtain different grammars. However, markedness is one step in the chain of explanation for: (a) the apparent step-wise variation of complexity and implicational universals of consonant clusters, (b) the Trubetzkoy hypothesis.
Pursuing the question of whether there are brain bases for phonological markedness necessitates an empirical and testable approach. This approach – put forward in this chapter – is couched in a theory of phonological feature specification and tested with brain imaging methods, with an emphasis on electrophysiology, encompassing electro- and magnetoencephalographic experiments. After sketching the theoretical background and introducing the basics of the reported brain measures, the chapter contains a review of some electrophysiological findings with their re-interpretation from the point of view of phonological markedness. Furthermore, the feature-approach is contrasted with a proposed reduction of phonological markedness to frequency-of-occurrence relations. Altogether, the review provides some important suggestions and starting points for further looking into the neurobiological reality of phonological markedness.
Markedness is a pervasive notion in theories of language. In some phonological theories or models, it very much occupies a prominent position. In this chapter it is argued that the notion of markedness is not useful to our understanding of phonology and language in general, for two reasons. The first is that the notion of markedness has convincingly been shown in the literature to be a confusing label for a variety of things, all of which can be explained independently. The second is that if phonology is to be understood as part of a biological system (which we call language), its components must be amenable to investigation in the cognitive and biological sciences. The notion of markedness as part of a phonological component does not seem to qualify for any meaningful kind of investigation in that respect. It follows that markedness has no place in biologically-informed phonology.
The true nature of “markedness” in the history of phonology is highly uncertain, in that the term is used to refer to a wide array of facts about language, and there is little agreement over what the term even refers to, much less whether it is a valid concept. This paper reviews certain applications of that concept in phonology, in search of some unity behind “markedness”. I show that “markedness” is about two unrelated things: formal properties of language, and functional probability of occurrence. Much effort has been put into forcing these two conceptions under a single computation umbrella, and that effort bears significant responsibility for the development of substance-dependent theories of grammar. As for whether “markedness” is a worthy topic of investigation, it is argued that the original formal question underlying markedness is still worth scrutiny in the theory of grammar: what is the nature of phonological features?
Halle’s (1959) argument against a distinction between morphophonemic and phonemic rules can be understood as an argument against the relevance of contrast to phonology. After adducing further arguments against a role for contrast, the paper provides a simple contrast-free analysis of the classic problem of the voicing behavior of Russian /v/. This segment undergoes voicing assimilation (like other obstruents), but does not trigger it (thus acting like the sonorants). In contrast to a long history of treating /v/ as a covert sonorant, the paper attributes the behavior of Russian /v/, which surfaces always as an obstruent, to underspecification with respect to the feature Voice.
The 21st century has brought major advances in our understanding of sound patterns, their phonetic basis, and their cross-linguistic diversity. Properties that were once thought to be universal, from phonological features to prosodic units like the syllable, have shown themselves to be highly variable across languages, and to sometimes fail at the very specific organizational role they are meant to play within sound systems. The goal of this chapter is to summarize evidence from the diversity of phonological systems that may inform areas of disagreement in modeling phonological grammars, with special attention to the locus of explanation in phonological theory. General arguments against phonological universals and phonological markedness in grammar are presented, including distinctive features, the sonority scale, and the prosodic hierarchy. Arguments for language-specific sound patterns and extra-grammatical explanations are also presented, including rare phoneme inventories, rare phonotactics, and distinct modality-specific properties of spoken versus signed languages.
Empty categories – positions in phonological representations that have no direct phonetic counterpart – are (still) controversial in phonological theory. In this paper we give the main arguments for assuming such positions and we furthermore establish a markedness hierarchy for empty positions: some of them are stronger (‘more marked’) than others, and we can derive this from a combination of Element Theory and Turbidity Theory. We illustrate our point with Italian and Dutch dialects, and point out that the phonological hierarchy of empty positions may correspond to a hierarchy of syntactic positions.
In terms of the roles of heads and dependents, there is a mismatch between phonology and syntax. In phonology, heads are important in both structural and informational terms. In syntax, on the other hand, heads have a different function: syntactic heads, like phonological heads, have an important structural role because they license dependent structure, but unlike phonological heads their informational role is relatively unimportant because they usually bear less linguistic (e.g. lexical) information than dependents. In order to achieve a greater degree of uniformity between phonology and syntax, this chapter proposes a reassessment of the roles of heads and dependents in phonology. Contrary to the widespread view, it is argued that heads in phonology, like those in syntax, are structurally important but lexically unimportant whereas dependents are structurally unimportant but lexically important. This view is supported not only by segmental distribution patterns but also by the size of the modulated carrier signal (rather than by the more standard phonetic measure, the sonority scale).
As a linguistic concept, markedness is in trouble (Hale & Reiss 2000, 2008; Blevins 2004; Haspelmath 2006; Samuels 2011; Reiss 2017). Accepting many of these criticisms and conclusions, this chapter, nevertheless, claims that there is space for markedness in formal phonology. By examining syllable structure from a highly representational viewpoint, I conclude that ‘markedness by complexity’ does have explanatory value and that markedness is explicable in terms of ‘structural complexity’ and ‘length of description’. The core demonstration is based on Charette’s (1990, 1991, 1992) formal typology of consonant clusters (CCs). These papers contain important implicational universals that Charette relates directly to her representations. My contribution will be to enrich the typology and reorganize those principles and parameters into a decision tree or parameter hierarchy that derives the implicational universals. Moreover, the number of parameter settings (the depth in the parameter hierarchy) increases the markedness of the resultant grammar. Crucially, each parametric ‘yes’ setting corresponds to an extra empty phonological category or extra ability to license, or reflects the distance between heads and dependents in the representation. For this reason, markedness is not merely a ‘metaphor’ ‘for a cognitive state’ (Haspelmath 2006), it is directly convertible into linguistic categories. This markedness is not part of the online computation of forms (contra Optimality Theory) and markedness statements cannot be re-ranked to obtain different grammars. However, markedness is one step in the chain of explanation for: (a) the apparent step-wise variation of complexity and implicational universals of consonant clusters, (b) the Trubetzkoy hypothesis.
Pursuing the question of whether there are brain bases for phonological markedness necessitates an empirical and testable approach. This approach – put forward in this chapter – is couched in a theory of phonological feature specification and tested with brain imaging methods, with an emphasis on electrophysiology, encompassing electro- and magnetoencephalographic experiments. After sketching the theoretical background and introducing the basics of the reported brain measures, the chapter contains a review of some electrophysiological findings with their re-interpretation from the point of view of phonological markedness. Furthermore, the feature-approach is contrasted with a proposed reduction of phonological markedness to frequency-of-occurrence relations. Altogether, the review provides some important suggestions and starting points for further looking into the neurobiological reality of phonological markedness.
Markedness is a pervasive notion in theories of language. In some phonological theories or models, it very much occupies a prominent position. In this chapter it is argued that the notion of markedness is not useful to our understanding of phonology and language in general, for two reasons. The first is that the notion of markedness has convincingly been shown in the literature to be a confusing label for a variety of things, all of which can be explained independently. The second is that if phonology is to be understood as part of a biological system (which we call language), its components must be amenable to investigation in the cognitive and biological sciences. The notion of markedness as part of a phonological component does not seem to qualify for any meaningful kind of investigation in that respect. It follows that markedness has no place in biologically-informed phonology.