An overview of the debate on the nature and origin of phonological features
1. Introduction
Phonology has been a central topic of inquiry in linguistic research for decades. Most modern phonological views assume speech sounds are not individual units but composed of multiple, interdependent properties called distinctive features. Distinctive features play a crucial role in linguistic sciences. They encode the patterns of possible speech sounds in languages worldwide and allow linguists to make robust predictions about the types of sound structures found in languages. The introduction of distinctive features is considered one of the most critical advances in linguistic science in the 20th century (Mielke & Hume, 2006)
However, despite the widespread acceptance of the existence of features, their nature and origin have been the subject of much debate in the linguistic community. There are currently three central claims about phonological features: first, that features are innately available and have nothing to do with phonetics (e.g. Hale & Reiss, 2008; Reiss & Volenec, 2022); second, that features do not exist as independent categories (e.g. Vihman & Croft, 2007; and Vihman, 2014); and third, that features result from language-specific learning, which is induced by pairings between acoustic dimensions of speech and their articulatory correlates (e.g. Stevens, 1989; Lin & Mielke, 2008; Mielke, 2008, 2011; Stevens & Keyser, 2010; and Dresher, 2014).
This paper aims to provide an overview of the debate on the nature and origin of phonological features by exploring each of the three claims in detail. The first part of the paper will examine the Phonological Perspective Denying Feature Existence, i.e. the No-Features View of phonology. The next part of the paper will explore views advocating phonological features: the Innateness Hypothesis and the Emergentist View. The final part of the paper will compare and evaluate the strengths and weaknesses of each perspective and discuss their implications for future research in phonology.
1.1. Phonological Features
According to Mielke and Hume (2006), the fundamental assumption of the modern phonological view is that speech sounds are not individual units but composed of multiple, interdependent properties called distinctive features. Phonological features are defined as the unique properties or attributes of speech sounds that distinguish the sounds from one another. In this view, features can be considered as the atoms of phonological systems. Features encode aspects such as place of articulation, manner of articulation, and voicing. The concept of distinctive features was introduced in the first half of the 20th century by The phonologists of the Prague School of Structural Linguistics who proposed that speech sounds are composed of a small set of binary features, such as [±consonantal] or [±voice].
According to Jakobson et al. (1952), it is possible to characterize any language using twelve distinctive features: vocalic/non-vocalic, consonantal/non-consonantal, interrupted/continuant, checked/unchecked, strident/mellow, voiced/unvoiced, compact/diffuse, grave/acute, flat/plain, sharp/plain, tense/lax, and nasal/oral. Distinctive features limit the number of possible speech sounds in languages worldwide and allow linguists to make powerful predictions about the types of sound structures found in language.
However, not all linguists agree on the exact number of distinctive features. Some linguists argue that there are more or fewer than the traditional twelve features, while others argue that certain features should be grouped or divided along diverse types of categories.
Above and beyond such debates, the literature opposes conflicting views regarding the nature and origin of features. The purpose of the present study is to investigate these divergent viewpoints, with each section of this paper offering a nuanced analysis of the various arguments put forth in the field.
1.2. Overview of the debate on the nature and origin of phonological features
The debate surrounding the nature and origin of phonological features has been ongoing for several decades. Part of this debate is about whether features are innate or learned. On the one hand, the Innateness Hypothesis (e.g., Hale and Reiss, 2002) posits that speech sounds are composed of a small set of binary features that are innate and present at birth. On the other hand, the No-Features View of Phonology argues that speech sounds are not composed of independent features but rather emerge through various contextual factors. Finally, the Emergentist View suggests that speech sounds are shaped by pairings between the acoustic dimensions of speech and their articulatory correlates, which must be acquired on a language-specific basis.
Despite these differing perspectives, each has evidence and arguments to support its claims. For example, the Innateness Hypothesis has been supported by language acquisition studies in infants, which suggest that certain speech sounds are processed differently, consistent with innate distinctive features (Kuhl, 2000). On the other hand, the No-Features View of Phonology has been supported by research on the articulatory development that shapes speech production behaviours (Vihman & Croft, 2007). Finally, the Emergentist View has been supported by research on the relationship between the acoustic and articulatory aspects of speech, which suggest that speech sounds are shaped by the interactions between these two factors (Lin & Mielke 2008; Mielke 2008, 2011; Stevens & Keyser 2010).
While each of these views has its strengths and weaknesses, they each contribute to our understanding of the nature and origin of phonological features. Furthermore, each of these views provides a unique perspective on the role of phonological features in language and how they contribute to the organization of speech sounds in language. In light of these differing perspectives, linguists must critically evaluate the evidence and arguments for each perspective.
1.3. Historical review of different ideas about the origin and nature of phonological features
The study of features as atomic units comprising speech sounds has evolved throughout history, with different perspectives emerging from pre-Structuralism, the Prague School of Structural Linguistics, American Structuralism, and Generative Phonology.[1]
In the late 19th century, Bell (1868) created a notation for speech sounds and developed a set of elements that could describe all possible sounds of the world’s languages. These elements are akin to what we now know as distinctive features. Sweet (1877) presented a descriptive system for vowels and consonants but did not examine the properties of phonemes.
In the early 20th century, the Prague School of Structural linguistics, with linguists such as Nikolai Trubetzkoy and Roman Jakobson, emphasized the use of contrastive features in phonological descriptions and proposed that features were elements of which speech sounds are made. Trubetzkoy suggested that a researcher had to pay attention to how a specific phonological system is patterned to arrive at a list of contrastive features. Jakobson proposed that the system of distinctive features was based on an ordered structure, leading to a divided inventory of features.
At the same time, American structuralists, such as Sapir and W. F. Twaddell, sought to find systematic patterns in language data and focused on the sound properties of language. Sapir (1925) proposed a distinction between phonemic and phonetic approaches to human language sounds but did not develop a theory of features. Twaddell (Twaddell, 1935) proposed a list of articulatory components of sounds and suggested that phonemes be defined in terms of them.
Universal Grammar (UG), as theorized by Noam Chomsky, has dramatically reshaped our comprehension of phonological features within the architecture of language (Chomsky, 1957). According to this theory, humans are innately equipped with certain linguistic abilities, enabling them to acquire language. Phonological features - the vocal distinctions, such as voicing, manner of articulation, and place of articulation - are considered part of the intrinsic building blocks of language structures as proposed by UG.
In UG, the Humboldtian concept of "making infinite use of finite means" is adopted (Humboldt, 1836). In terms of phonology, this means that the finite set of phonological features can be recombined in infinite ways, yielding a myriad of sound patterns across various languages.
UG also suggests that grammar includes rules that map abstract structures onto specific sound representations - elements that supposedly belong to a 'universal phonetics'. This extends the influence of UG beyond syntax and morphology into phonetics and phonology, and provides a holistic framework for language structure and acquisition (Chomsky, 1965).
The implications of UG for phonological features is most evidently demonstrated in the collaboration of Morris Halle and Noam Chomsky in the book "The Sound Pattern of English" (SPE). The authors built on Halle's earlier work, "The Sound Pattern of Russian" (Halle, 1959), and combined it with Chomsky's UG theory to propose a model for the phonology of English.
Halle's work on Russian phonology sought to establish a generative phonological theory, highlighting the importance of rule-based derivations and transformations in sound patterns (Halle, 1959). This resonated with the generative grammar model of UG that Chomsky had proposed.
In the "Sound Pattern of English," Halle and Chomsky demonstrated how their combined understanding could be applied to English. They developed a system in which phonemes were represented as a bundle of distinctive features, and phonological rules operated on these features to generate different phonetic realizations (Chomsky & Halle, 1968). Their model emphasized the systematic and rule-governed nature of phonology, providing a foundation for subsequent research in phonological theory.
The significant contribution of UG to the understanding of phonological features, as exemplified by the work of Halle and Chomsky, highlights how our understanding of the sound structures of languages is intimately tied to our inherent linguistic abilities, reinforcing the universal nature of language structure and acquisition.
Following the publication of The Sound Pattern of English,” many phonologists attempted to address the view’s shortcomings. Autosegmental Phonology (Goldsmith, 1976) and Optimality View (Prince & Smolensky, 2004) emerged as offshoots of Generative Phonology, and features were also considered innate in these theories.
The evolution of understanding phonological features has bridged various theoretical approaches, each contributing to the development of Generative Phonology within the framework of Chomsky's Universal Grammar. The groundbreaking collaboration between Halle and Chomsky, evidenced in "The Sound Pattern of English," exemplifies this comprehensive understanding. Their work underscored the significance of phonological features as part of our inherent linguistic capabilities, linking them with other linguistic phenomena. As we continue to explore this realm, our comprehension of these crucial structures will likely grow, fueling further advancements in the field.
2. The Phonological Perspective Denying Feature Existence
The No-Features View of Phonology proposes an approach to phonology that challenges the traditional view of phonology consisting of a set of abstract and discrete features. Instead, this view suggests that phonology emerges due to the cumulative effects of phonetic constraints derived from the physical properties of speech sounds and articulatory gestures modulated by the speaker’s goals and communicative context.
This view emphasizes the role of phonetic substance in shaping the phonological organization with interactions among phonetic constraints giving rise to phonological patterns that vary along a continuum.
This alternative approach has significant implications for our understanding of phonological development. According to the No-Features View of Phonology, children’s phonological systems are not structured around innate, pre-determined features but are shaped by their experience with the speech sounds in their language. This perspective directly challenges the notion of a universal set of phonological units that underly all human languages. The absence of features in the radical templatic model calls into question the idea of language-independent, pre-specified feature sets as the building blocks of phonological systems.
This templatic approach to phonology gains further reinforcement from non-linear representations. According to this approach, phonological features are not necessarily confined to particular segment positions within a word but can stretch across several positions, such as solely consonantal or vocalic slots. The mapping of features onto skeletal positions is realized by allocating each feature to a unique tier. Articulatory phonology pushes this tendency even more by adopting articulatory gestures as fundamental building blocks and non-linearly mapping them to produce words. As Vihman and Croft (2007, p. 717) point out, Articulatory phonology is a directly phonetically based non-linear model, in which the articulatory gestures are the basic phonological ‘‘features’’, and the non-linear mapping of gestures is the result of the complex motor coordination of the gestures to produce a word. The execution and coordination of articulatory gestures are the sources of most phonological processes.”
A vital aspect of the No-Features View of phonology is the rejection of the idea that children start with a set of innate or pre-existing phonological features that they gradually learn to apply to speech sounds. Instead, phonological organization is viewed as a property that arises from the interaction between the speaker’s developing articulatory and perceptual capacities and the phonetic substance of the input.
According to the No-Features View of phonology, the emergence of phonological organization is a gradual and continuous process shaped by multiple factors, including the child’s experience with the ambient language, the distributional properties of the input, and the child’s own articulatory and perceptual abilities. There is no clear point at which the child can be said to have phonology as a separate, fully formed system. The following represents what Vihman and Croft (2007) say about this view:
It proposes that a limited number of specific, actual word shapes are the first steps in phonological learning. The child gradually develops one or a small number of phonological templates, then a wider variety of them, while at the same time inducing a range of other phonological categories and structures from the known word shapes. The result of differentiating and generalizing knowledge of the phonological structure of words in the course of language acquisition is an adult template-based model of phonological representation, with neither discontinuity nor an assumption of pre-specified adult competence. (p. 686)
This view challenges the traditional approach to phonology, which relies heavily on the concept of features as the basic building blocks of phonological systems. Instead, it suggests that phonological structure results from a complex interaction between input, perceptual and articulatory abilities, and the cognitive mechanisms that process and organize that input. Phonology emerges as the child gains experience with the sounds of their language and begins to abstract away from the raw data. Vihman (2014, p. 315) notes that It makes sense to assume that advances in lexical knowledge and experience will facilitate further learning (over and above the effects of age or maturation), and there is theory as well as a certain amount of empirical data to support that supposition.”
According to the Radical” Templatic model of phonology proposed by Vihamn and Croft (2007), segmental phonological structure is represented as language-specific phonotactic templates acquired through a usage-based abstraction process based on babbling practice and input experience. These templates reconcile the child’s phonetic repertoire with adult phonological patterns. In adult phonology, segment categories are best defined in terms of their occurrence in positions in the templates of individual languages. Pervasive variation in phonetic realization and phonological distribution patterns provides cross-linguistic evidence supporting the template construct.
Critiques of feature-based phonological theories have been widely discussed in the literature. One of the main criticisms is that the concept of phonological features is too abstract and arbitrary and does not adequately capture the nature of phonological representation. According to this view, feature-based phonology is prematurely abstracted from the empirical data and does not correspond to how children learn and use language.[M1]
Another critique of feature-based phonology is that it cannot account for the full range of phonological phenomena observed in natural languages. In particular, it has been argued that features cannot capture the non-linear and non-local aspects of phonological structure, such as phonotactics and stress patterns. Menn and Vihman (2011) propose that phonology may emerge from the interaction of multiple sources of information, including distributional learning, phonetic detail, and articulatory constraints:[M2]
We see each child as gaining knowledge, first, from distributional learning of many kinds in the period before the production of the first (recognized) words, and then additionally, in the period that we have focused on here, from the words s/he is using, out of which both units and patterns are gradually induced. (p. 282)
It has also been argued that feature-based phonology does not explain how children acquire the specific features of their native language (Vihman & Velleman, 2000). According to Vihman and Velleman (2000), children may initially use a holistic, whole-word approach to phonology, gradually breaking down words into smaller units over time. They suggest that children’s early phonological representations are based on a combination of auditory, motor, and social information.
The role of the lexicon in phonological organization is a central theme in the No-Features View of phonology. In this view, phonological patterns are not derived from abstract phonological features. Instead, they are built up through the child’s exposure to and learning of the sound patterns of the language. As Vihman and Croft (2007) put it:
we see the earliest phonological organization as constituting an inductive generalization based on the child’s first repertoire of phonetic patterns and their interaction with the phonological structure implicit in the words of the ambient language that the child is attempting to reproduce. The phonological organization itself inheres in whole word patterns or word templates, as can be seen from the adapted patterns illustrated above. Phonological categories will gradually emerge later, in different ways for different children. (p. 708)
One key aspect of this approach is that the child’s early lexicon is not organized according to abstract phonological features but based on acoustic-phonetic similarity. Vihman (2010, p. 264) notes that children’s early words tend to be phonetically similar: a child’s word forms may be more similar as a set than they are like their adult models on a word-by-word basis.” This view suggests that this organization of the lexicon provides a foundation for developing phonological patterns. It is from this word-based view that templates make sense as generalizations.
This approach is consistent with findings from research on the role of experience in phonological learning. Vihman claims that experience shapes perception and production, and this shaping continues throughout the lifespan. After mentioning advances in the child’s neuromotor control, attentional mechanisms, and perceptual capacities, Vihman (2014, p. 2) claims these advances build on the child’s emergent pragmatic and phonetic skills and social experiences to complete the set of essential precursors to language use.” This suggests that the phonological patterns of a language are not fixed or pre-determined but instead are constantly evolving based on the experiences of the language speakers.
Findings from cross-linguistic research on early word production support this perspective. Vihman (2014) notes that children’s early words often contain non-native sounds, suggesting that they are not simply memorizing a set of discrete segments and their features. Instead, she argues that children use prosodic templates to generate words and that these templates can be adapted to new sound patterns as the child’s phonetic inventory expands. This view is consistent with the No-Features View, which proposes that the phonological system is based on units larger than individual segments and that these units are learned through experience with the sound patterns of the child’s language.
The No-Features View of Phonology challenges the traditional belief of phonology as consisting of a set of abstract and discrete features. Instead, it suggests that phonology emerges due to the cumulative effects of phonetic constraints derived from the physical properties of speech sounds and articulatory gestures modulated by the speaker’s goals and communicative context.
This approach has significant implications for our understanding of phonological development. It rejects the notion of a universal phonological grammar underlying all human languages. It posits that phonological organization arises from the interaction between the speaker’s developing articulatory and perceptual capacities and the phonetic substance of the input. While the concept of features has been widely criticized, the No-Features View of Phonology offers a different perspective on the nature of the phonological structure. It highlights the vital role of the child’s experience in shaping their phonological system.
3. Views Advocating for Phonological Features
The existence and significance of phonological features are generally acknowledged by linguists, given their pivotal role in facilitating comprehension and use of language. Despite this agreement, considerable divergences exist, particularly when it comes to understanding the origin and nature of these features. This presents two principal theories: the Innateness Hypothesis and the Emergentist View. While both theories agree on the existence of phonological features, they provide distinct interpretations of their sources and manifestations within the human linguistic system.
The Innateness Hypothesis presents the first perspective, postulating that phonological features are innate and pre-determined. On the other hand, the Emergentist View proposes a contrasting view, positing that phonological features emerge as a product of perceptual and articulatory interactions during language acquisition. These theories, explored in the subsequent sections in greater depth, offer comprehensive overviews of their central tenets and supporting evidence. Through the detailed discussions that follow, a more nuanced understanding of the theories advocating for phonological features is developed, illuminating the ongoing debates about the nature and origin of these features.
3.1. The Innateness Hypothesis
3.1.1 The Nature of Phonological Features
The Innateness Hypothesis suggests that these features are inherent, biologically ingrained aspects of our cognitive function rather than elements learned through interaction with and exposure to a specific language. This notion that phonological features are innate and thus biologically determined was explicitly put forth by scholars such as Hale and Reiss (2008). They proposed that phonological features are a form of cognitive primitives. In cognitive psychology, primitives are viewed as the most fundamental elements of cognitive systems. These cognitive primitives are not derived or learned from data or external sources but are inherent to the cognitive system itself. They exist as mental entities, independent of any specific language or culture, and are universally present across the human population, constituting a part of our innate cognitive endowment. Consequently, in phonology, cognitive primitives form the foundation for more complex linguistic structures and processes. These universally shared entities provide a lens through which to understand the innateness of phonological features.
Following this line of thought, Reiss and Volenec (2022) further advanced the argument for the innateness of phonological features by suggesting that these elements are innate and substance-free.” Here, they introduced an element of abstraction to characterizing phonological features, taking a departure from traditional perspectives. By the term substance-free,” Reiss and Volenec implied that phonological features are devoid of direct articulatory and acoustic content or have no explicit reference to such phenomena. In their view, phonological features are not reliant on the physical sounds and articulatory actions that create speech. Instead, they are abstract cognitive entities represented symbolically in the brain.
Phonology is one part, or one module, in technical terms, of that cognitive, brain-based faculty. This module consists of computational and representational aspects. The computations are ordered logical operations, rules, and they apply in a manner that is blind to phonetic substance. The representations are made from elementary units, features, which are symbols in the brain, in the sense of Gallistel and King (2009). These symbols are innate (this is nativism) – they are knowledge unlearned and untaught” (Halle 1978), which is bestowed by human biology. They are also devoid of phonetic substance. Thus, neither the computational nor the representational aspect of phonology contains phonetic substance (this is formalism), which is why this approach is called Substance Free Phonology. (p. 606)
This novel interpretation of phonological features as substance-free cognitive symbols housed within the brain is rooted in the formalist approaches to phonology. These formalist approaches propose a complete abstraction of phonological features from the physical, phonetic substance of speech. They argue that phonological features are abstract mental entities functioning within a computational system that is blind to the physical, phonetic substance of speech. As such, these features are seen as having only an indirect relation to phonetic substance; a notion echoed in Reiss and Volenec’s work.
The perception of phonological features as substance-free symbols represented within the brain has been further corroborated by empirical evidence from various studies. One such study was conducted by Pulvermüller et al. in 2006, employing brain imaging technologies. In this study, participants were presented with both words and non-words that shared the same phonological features. The brain imaging results indicated that the same brain areas were activated when participants were presented with words and non-words, suggesting that the processing of phonological features can occur independently of semantic meaning. This study’s findings underscore the idea that phonological features are abstract mental entities processed independently of the physical acoustic or articulatory properties of speech sounds. They thereby challenge the traditional assumptions within the field of phonology about the inherent and direct connection between phonetic substance and phonological features.
Further building on this, Reiss and Volenec argue that a crucial step in understanding the nature of phonological features is acknowledging their independence from phonetic substance. This shift in perspective holds significant implications for our understanding of language acquisition. They propose that an innate phonological system, replete with a set of universal and substance-free features, equips infants with the necessary tools to acquire language, regardless of the phonetic variability among different languages.
Their theory presents an abstract and substance-free character of phonological features. This intriguing viewpoint provides an intriguing explanation for the remarkable ease and speed with which humans acquire language. Despite the considerable variation in the sounds and articulatory mechanisms of the world’s languages, humans can rapidly and effectively learn any language. This suggests that the acquisition process depends not on phonetic substance but on abstract, cognitive representations of phonological features. This conceptualization of phonological features as abstract, substance-free, and innate mental entities reshapes our understanding of the nature of these elements and the cognitive processes underlying speech and language.
3.1.2. The Origin of Phonological Features
Concerning the origins of phonological features, the Innateness Hypothesis argues that these features are biologically determined and innate, not learned through exposure to a particular language (Hale & Reiss, 2008). This hypothesis challenges the traditional view that phonological features are derived from sensory experiences and conditioned by specific language environments.
Reiss and Volenec (2022) support the Innateness Hypothesis, maintaining that phonological features are substance-free symbols in the brain and do not rely on phonetic substance. In their conceptual framework, these features are universal cognitive entities that exist independently of specific languages. They argue that phonology forms a segment of the cognitive system or, more accurately, a module encompassing computational and representational components. The computations involve logically ordered operations, or rules, which are blind to phonetic substance, while the representations consist of basic units, and features, that act as symbols in the brain. These symbols are innately endowed, unlearned, and untaught (Halle, 1978), gifts from human biology, and void of phonetic substance.
Gallistel and King’s (2009) idea of cognitive primitives as the essential elements of human cognition present at birth aligns perfectly with this conceptualization of phonological features. These primitives represent the innate cognitive structures that constitute the bedrock of intricate cognitive processes. They are universally shared entities unaffected by language or cultural variations, contributing to the premise that phonological features are innate and substance-free.
[T]he number of symbols that might have to be realized in a representational system with any real power is for all practical purposes infinite; it vastly exceeds the number of elementary particles in the universe, which is roughly 1085 (or 2285), give or take a few orders of magnitude. This is an example of the difference between the infinitude of the possible and the finitude of the actual, a distinction of enormous importance in the design of practical computing machines. A computing machine can only have a finite number of actual symbols in it, but it must be so constructed that the set of possible symbols from which those actual symbols come is essentially infinite. (By ‘essentially infinite’ we will always mean greater than the number of elementary particles in the universe; in other words, not physically realizable.) This means that the machine cannot come with all of the symbols it will ever need already formed. It must be able to construct them as it needs them – as it encounters new referents. (Gallistel & King 2009: 74)
The existence of cognitive primitives is evidenced by infants’ natural ability to distinguish between speech sounds of any language, regardless of prior exposure (Kuhl, 2004). This ability indicates the presence of a universal set of phonological primitives that form part of the human cognitive endowment. Cognitive primitives also explain the universality of grammatical categories, such as nouns and verbs across languages, reflecting the human cognitive capacity to categorize and understand the world.
The concept of phonological features as substance-free symbols is supported by neuroscientific research. For instance, Pulvermüller and colleagues (2006) demonstrated that listening to words or non-words with similar phonological features activated similar brain areas associated with language processing, irrespective of their semantic content. This finding implies that the brain processes phonological features independently of their phonetic substance, challenging traditional phonological assumptions that tie phonetic substance and phonological features closely together.
We can better understand language acquisition by seeing phonological features as independent from phonetic substance. Reiss and Volenec (2022) propose that an innate phonological system equipped with a universal feature set enables infants to acquire language, regardless of phonetic variability across languages. The abstract and substance-free nature of phonological features illuminates the human mind’s ability to learn languages with remarkable ease and speed amidst vast phonetic diversity.
Chomsky’s (1981) theory of Universal Grammar (UG), a set of innate linguistic principles and parameters, lends further credence to the Innateness Hypothesis. UG is considered a cognitive framework that enables the acquisition and use of language, including phonological features, suggesting that these features are biologically determined, not learned.
درباره این سایت