Showing papers on "Naturalness published in 2002"

Evolution and Media Naturalness: A Look at E-Communication Through a Darwinian Theoretical Lens

Abstract: Modern theories of human evolution converge on the belief that our brain has been designed to cope with problems that occurred intermittently in our evolutionary past. Evidence suggests that, during over 99 percent of the evolutionary process leading to the emergence of our species, our ancestors communicated in a synchronous and colocated manner, and employing facial expressions, body language, and oral speech (what we refer to here, generally, as “face-to-face” communication). Thus, it is plausible to assume that many of the evolutionary adaptations our brain has undergone in connection with communication have been directed at improving the efficiency and effectiveness of face-to-face communication, which begs the question: What happens when we selectively suppress face-to-face communication elements (e.g., colocation, the ability to employ/observe facial expressions) through e-communication technologies? This paper tries to provide an answer to this question by developing a hypothesis, called the media naturalness hypothesis, which builds on modern human evolution theory. The media naturalness hypothesis argues that, other things being equal, a decrease in the degree of naturalness of a communication medium (or its degree of similarity to the face-to-face medium) leads to the following effects in connection with a communication interaction: (1) increased cognitive effort, (2) increased communication ambiguity, and (3) decreased physiological arousal. It is argued that the media naturalness hypothesis has important implications for the selection, use, and deployment of e-communication tools in organizations, particularly in the context of business-to-consumer interactions.

Naturalness: Beyond Animal Welfare

Abstract: There is an ongoing debate in animalethics on the meaning and scope of animalwelfare. In certain broader views, leading anatural life through the development of naturalcapabilities is also headed under the conceptof animal welfare. I argue that a concern forthe development of natural capabilities of ananimal such as expressed when living freelyshould be distinguished from the preservationof the naturalness of its behavior andappearance. However, it is not always clearwhere a plea for natural living changes overinto a plea for the preservation of theirnaturalness or wildness. In the first part ofthis article, I examine to what extent theconcerns for natural living meet ``theexperience requirement.'' I conclude that someof these concerns go beyond welfare. In thesecond part of the article. I ask whether wehave moral reasons to respect concernsfor the naturalness of an animal's living thattranscend its welfare. I argue that the moralrelevance of such considerations can be graspedwhen we see animals as entities bearingnon-moral intrinsic values. In my view the``natural'' appearance and behavior of an animalmay embody intrinsic values. Caring for ananimal's naturalness should then be understoodas caring for such intrinsic values. Intrinsicvalues provide moral reasons for action iffthey are seen as constitutive of the good lifefor humans. I conclude by reinterpreting,within the framework of a perfectionist ethicaltheory, the notion of indirect dutiesregarding animals, which go beyond andsupplement the direct duties towardsanimals.

Naturalness bounds on dipole moments from new physics.

Abstract: Assuming naturalness that the quantum corrections to the mass should not exceed the order of the observed mass, we derive and apply model-independent bounds on the anomalous magnetic moments and electric dipole moments of leptons and quarks due to new physics.

Factors influencing ratings of speech naturalness in augmentative and alternative communication

Abstract: The concept of speech naturalness has been used in the field of speech-language pathology as a clinical measure of perceptual quality of “normal” and “not normal” speech. Whereas measures of intelligibility have been commonly used to assess the quality of voice output augmentative and alternative communication (AAC) devices using DECTalk™ speech, measures of speech naturalness have not. Three studies were conducted to determine the effects of manipulation of rate, pitch, and pause on ratings of speech naturalness by naive listeners of DECTalk synthetic speech. The results indicate that DECTalk speech characterized by faster rate and no added pauses was perceived as being more natural than speech with slow rate and added pauses. Manipulation of pitch had no effect on naturalness ratings.

33.1: Invited Paper: Image Quality is FUN: Reflections on Fidelity, Usefulness and Naturalness

Abstract: Differences in perceived naturalness and image quality have led to an information-processing approach towards understanding image quality, thus challenging the traditional, signal-processing approach. This paper presents a general image quality model integrating both approaches. The model assumes the existence of three constraints, i.e. fidelity, usefulness and naturalness. Image quality is modeled as the weighted sum of these constraints whereby the weighting depends on task, context, image content etc. Model predictions will be confronted with experimental data on optimizing color reproductions of natural scenes.

Gold, jade, and emeruby: The value of naturalness for theories of concepts and categories.

Abstract: Researchers studying the psychology of concepts frequently draw distinctions between artificial and natural concepts. Unfortunately, there is a lack of consensus regarding the foundations and implications of the distinction. This paper provides a review and evaluation of the different ways researchers have approached the question of conceptual naturalness Accounts may be divided into two approaches described as psychologically or externally based. These characterizations motivate distinctive sets of research questions. In addition to the particular implications, I also consider the general significance of a distinction between natural and artificial concepts. Natural Categories 3 Gold, Jade, And Emeruby: The Value Of Naturalness For Theories Of Concepts And Categories Under the rubric of post-modernism it is possible to read all manner of challenges to commonsense intuitions about the natural order of things. Whether focusing on particular distinctions we make (such as racial or gender groupings), or on our entire system of beliefs about the kinds of things there are in the world, the argument often goes that we are misguided in thinking that our concepts are products or reflections of an objective nature (e.g., Fish, 1996; Rorty, 1999, cf. Hacking, 1999). There is nothing natural or inevitable about any particular system of concepts. This question seems a basic issue for scientific research on concepts. One of the goals of Psychology and Cognitive Science is to answer the questions of which (if any) concepts are natural and why. Unfortunately, researchers use the designations “natural” and “artificial” in different ways. It seems that the field has no unequivocal answer to the question of which concepts are natural. As a consequence there is no consensus regarding the significance of the distinction for theories of concepts and categories. The goal of this essay is to review the different senses in which concepts may be deemed natural and to draw out the research and theoretical implications of different senses of naturalness. Early researchers often employed meaningless stimuli in concept learning experiments (e.g., Hull, 1920). A typical experiment required participants to learn a concept such as SMALL OR BLACK (e.g., Goodnow, Bruner, & Austin, 1957). Although these procedures were occasionally criticized as artificial and non-representative of the actual process of concept acquisition (Kakise, 1911; Smoke, 1932) the assumption was that the concepts derived did not differ in meaningful ways from concepts derived outside the experimental setting. In the 1970’s Rosch (1973, 1978) posed two, linked, challenges to this view. First, she argued the materials were unrepresentative of naturally occurring objects. Second, the psychological processes involved in categorizing such materials were not the ones typically used in forming concpets. Rosch's theory involves claims about psychological naturalness and external naturalness. The human mind is disposed toward concepts with a family resemblance structure; construction and use of such concepts is natural. But such dispositions are no accident. Objects in the world actually do form clusters around correlated attributes; the groups are natural. In Rosch's theory it is possible to discern two senses of conceptual naturalness as well as a hypothesis about their Natural Categories 4 interrelation. Moreover, the research implications also seem clear. It is natural concepts that should be the focus of research. The psychology of concepts is advanced by studies of how people form and use natural concepts, such as BIRD, rather than by studies of artificial concepts, such as SMALL OR BLACK. In the remainder of this review I will discuss additional psychological and external accounts of naturalism as well as further suggestions about their interrelations and the implications for the psychological study of concepts. PSYCHOLOGICAL CRITERIA FOR CONCEPTUAL NATURALNESS Much psychological work begins with the intuition that some concepts are more psychologically natural than others. The idea is that we tend to form some concepts rather than others. For example, Osherson (1978) characterizes natural concepts as those that are acquired during the course of normal experiences. On this perspective, concepts are orderings imposed on experience by the mind. Cognitive principles determine which concepts are formed (how experience is mapped onto concepts). Those principles define the set of concepts natural within a given cognitive system. The operation of the principles of conceptualization will have consequences indicative of naturalness, for example some concepts will be more easily learned than others. Ease of acquisition may be only one of a number of effects of naturalness. A theory of concepts will describe the set of cognitive principles responsible for those effects, those processes that make some concepts more natural than others. Within the psychological literature a number of such principles have been proposed. Structural constraints The strongest psychological formulations of naturalness involve claims that there are structural constraints on the types of concepts which may be represented. Natural concepts have some formal properties that artificial ones lack. Structural constraints are generally thought to derive from the format with which concepts are represented. Thus if concepts are represented in a propositional form, artificial concepts would include those involving logical contradictions (e.g., COLORLESS GREEN STONE). Similarly, the representation of spatial knowledge may be incommensurate with some concepts (e.g., 6 DIMENSIONAL CUBE). A specific proposal along these lines is the suggestion that natural concepts fit into a hierarchical organization of ontological Natural Categories 5 types (Keil, 1979, 1981; Osherson, 1978). A consequence of this model is that the predicates applicable to a concept must apply to all concepts below it in the hierarchy (its daughters). If weight is a predicate of OBJECT, then it is also a predicate of all concepts below OBJECT (kinds of OBJECT such as LIVING THING). Concepts involving cycles in the ontological tree structure are unnatural. For example, objects and waves seem to be ontologically different sorts of concepts. Concepts of objects accept predicates of mass and weight but not of frequency. Concepts of waves accept predicates of frequency but not mass. A concept for which both sorts of predicates were appropriate (e.g., ELECTRON) would be unnatural. Either all objects have a frequency or none do (see Carey, 1986 for other counter-examples). Representational format constraints are often involved in developmental claims. For example, the concepts formed by young children may be limited to those with structures that are: complexive (vs definitional, Bruner, et al., 1956; Keil & Batterman, 1984; Vygotsky, 1962), wholistic (vs. analytic, L. Smith, 1989), or organized in a single level (vs. taxonomically organized, see Johnson, Scott, & Mervis, 1997). In the developmental cases the constraints are often (but not always) described as absolute; it is impossible for children to form unnatural concepts. In contrast, adults do seem able to construct unnatural concepts (at least those of the sorts given as examples above). That people use and form unnatural concepts is a challenge to theories of structural constraints. One way to meet the challenge is to posit multiple representational systems with different structural constraints. For example, mathematical or linguistic representational systems may generate concepts violating natural principles. A capacity to think mathematically may allow formation of concepts that could not be represented otherwise. Such 'externally' generated concepts would always retain their unnatural character, perhaps never completely integrated with other, natural, concepts (unless representational formats change, see Carey & Spelke, 1993). This line of reasoning raises the possibility that people may have multiple sets of incommensurable concepts. Concepts violating natural principles need not be completely excluded from cognition. Nonetheless, this relatively strong construal of naturalness does suggest that unnatural concepts will be rare and formed only under exceptional circumstances (see Sperber, 1990; Osherson, 1978). Note that the converse does not necessarily hold: There are likely many concepts satisfying structural constraints on naturalness that are not commonly acquired in the course of experience Natural Categories 6 (e.g., EMERUBY, an emerald examined before time T or a ruby thereafter, Goodman, 1955). Thus structural constraints may provide necessary but not sufficient criteria for naturalness. Processing principles An alternative perspective on psychological naturalness characterizes the property in terms of optimality rather than as accordance with constraints. A concept either does or does not violate some structural constraint. In contrast, many psychological characterizations present naturalness as a matter of degree; concepts are more or less natural relative to each other. In such accounts naturalness is a consequence of the operation of psychological processes of concept formation and use. For example, one straightforward determinant of naturalness may be familiarity. The more frequently people use a concept to make inferences, the more natural it seems (see Goodman, 1955). Thus while any concept involving a feature such as NON-RED would seem artificial, familiarity could lead people to accept the concept SAPPHIRE (defined as NON-RED CORUNDUM). It is also possible to reconceptualize the structural constraints described above as matters of degree or preference. For example, some cycles in the ontological tree may be less problematic than others (Kelly & Keil, 1985). In practice it may be difficult to distinguish real structural constraints from strong preferences (see Kalish, 1998; Keil, 198

The naturalness of object orientation: beating a dead horse?

Abstract: I n a recent column (\" A Story about the Creativity Involved in Software Work, \" Sept./Oct. 2001), I hinted at a research study whose goal was to compare the object-oriented, function/process, and information/data approaches to systems analysis. I called it \" a story for another time. \" This is that time. In that column, I described my boredom in looking at videotapes of graduate students analyzing an assigned information systems task. My point was that what at first appeared to be boring activity turned out to be thinking activity, and that the thinking part of the task overwhelmed the task's more clerical aspects at an 80-to-20-percent ratio. That was strong evidence, I suggested , that software development is far more complex than those who believe it is easy and automatable have led us to believe. That was that column; this is now. I want to revisit this topic to explain where those \" boring \" videotapes came from and why their original purpose was at least as important as the purpose I derived from them. Amazing benefits! The software field has been subjected, over the years, to excessive claims of benefits for almost every new technology. Fourth-generation languages were to lead to \" programming without programmers, \" CASE tools would bring about \" the automation of programming , \" and object orientation was to be a dramatic new methodological approach to systems development that would replace all the other, older, methodologies. Two phenomena have accompanied such excessive claims: I Once the concepts are more thoroughly understood, the benefits turn out to be far more modest than claimed. I That transition from excessive claims to modest benefits has seldom been accomplished with the aid of evaluative research. Researchers, I am saying here, have been of little help in sorting the technological and methodological wheat from its chaff. That's why I am so impressed by the study I hinted at in the previous column. Videotape evidence! This study was an honest-to-goodness attempt to objectively compare the benefits of the (at that time) newfangled object-oriented approaches with the more traditional func-tion/process and information/data alternatives. (That the study's lead researcher, Iris Vessey of Penn State University [and now of Indiana University], is my wife also partly accounts for my being impressed!) The claims made for object orientation, then as now, were that it is a more natural form of problem solution and that …

Rule naturalness and the acquisition of phonology

Abstract: Standard generative phonology: privileged status to " natural " phonological rules (those with a clear phonetic motivation). Making it easier for theories to express common processes (which tend strongly to belong to the natural category) is often considered a primary goal. " Unnatural " phenomena are typically relegated to a " peripheral " or secondary status. The usual (tacit) assumption: they require some other apparatus in the grammar. If nothing else, the child ought to first attempt to apply basic principles of the phonology — biased toward naturalness — and only when this fails will she resort to secondary principles. An implication of this sort of approach — implicit or explicit — is that a child learner equipped with the universal phonological faculty will more easily learn processes that are natural. (1) Morphophonological patterns that are in accord with natural phonological predispositions will enter early and will seldom lead to errors. (2) Phonological rules with a heavy natural component (e.g. voicing assimilation) appear to be acquired early. Problem: If children are biased toward natural processes, processes ought to become more natural over time, or at least not become less natural. Language transmission is imperfect, and bias toward natural processes should err in favor of them. But phonological rules tend to become less natural over time (cf. Hyman 1975: 181f): (3) Although sound changes are sometimes blocked by considerations within a paradigm [...] no corresponding force has been discovered which would strive to keep rules natural. Instead, the above examples show the great tendency for rules to become unnatural [...] that is, to lose their phonetic plausibility and become morphologically conditioned. Natural and unnatural rules in the same language can have similar status. Kashaya (Pomoan: N. California) has lowering to [a] after a uvular (natural); and /i/ becomes [u] after [d] (unnatural, and arose out of a morphological reanalysis: Buckley 2000).

The effect of variation in naturalness on phonetic perceptual identification

Abstract: The relation between apparent naturalness and phonetic identification was assessed in six perceptual tests. A seven‐step place‐of‐articulation series spanning [da] to [ga] was created with speech synthesis approximating the spectra of natural samples. The sensitivity of perceivers to this realization of a place contrast was assessed by estimating the cumulative d′ across the series in identification tests. Four variants of this series differing in apparent naturalness were produced by altering the synthesis source function while preserving the center frequency and bandwidth of the formants, and by replicating the gross spectrotemporal patterns with time‐varying sinusoids. In addition to calibrating perceivers’ sensitivity to the place contrast over variation in naturalness, we conducted a naturalness tournament composed of items drawn from the five test series. A correlation of the findings of the naturalness tournament with the measures of phonetic sensitivity offers an index of the effect of variation i...

Speech unit selection based on target values driven by speech data in concatenative speech synthesis

Abstract: In concatenative speech synthesis systems, speech features (prosody, etc.) are estimated as "target values" according to the input text at the start of processing. The target values are used to evaluate each speech unit in the speech database of the system in order to find out a "nearest" unit, which shows the lowest cost, to the value. In conventional systems, the target value is treated as an absolute value to calculate the cost of each unit. In this paper, we propose a new method to drive the target value depending on the features of speech units in the synthesis system within the range in which naturalness and clarity are maintained. This method was applied to fundamental frequency control, which is one of the most important factors in maintaining naturalness of synthesized speech, based on a superpositional model for a 10-name (Japanese place names) synthesis test, and it was found that the cost decreased an average of 22.9%.

Naturalness in english: some (morpho)syntactic examples

Abstract: In Slovenia, the natural syntax of the Klagenfurt brand has been extended to the study of the behav­ iour of (near-)synonymous syntactic expressions, here called syntactic variants. The work below is illustrated with (morpho)syntactic cases from English. (Naturalness Theory applied to English has sofar not received much attention.) About a half of the examples deal with relative clauses; the other half considers fronting phenomena. The language material is divided into consecutively numbered deductions, in each of which the existence of a (morpho)syntactic state of affairs ispredicted on the basis of apposite assumptions and Andersen ’ s markedness agreement rules. The subject-matter of this paper is a (language-universal) theory developedin Slovenia by a small group of linguists (under my guidance) that mainly useEnglish, German, and Slovenian language material as the base of verification. Our workowes much to, and exploits, the (linguistic) Naturalness Theory as elaborated especially at some Austrian and German universities; cf. Mayerthaler (1981), Wurzel(1984), Dressler et al. (1987) and Dressler (2000). Naturalness Theory has also been appliedto syntax, notably at the University of Klagenfurt; the basic references are Dotter 1990), Mayerthaler & Fliedl (1993) and Mayerthaler et al. (1993; 1995; 1998). Within thenat uralsyntaxoftheKlagenfurtbrand,theSlovenianworkgrouphasbuiltanextension that studies the behaviour of (near-)synonymous syntactic expressions, here calledsyn­ tactic variants. Whenever two syntactic variants are included in the samenaturalness scale, and consequently one variant can be asserted to be more natural than theother, something can be said about some grammatical properties of the twovariants.

Systematic downgrading for investigating ‘‘naturalness’’ in synthesized singing using STRAIGHT: A high quality VOCODER

Abstract: Conceptual simplicity of the classical channel VOCODER provides a powerful means for systematic investigations on perceptual effects of speech related physical parameters when combined with modern computational power and signal processing theories. A modern version of channel VOCODER, STRAIGHT [Kawahara et al., Speech Commun. 27, 187–207 (1999)], which is also an extension to pitch‐synchronous analysis and synthesis, generates naturally sounding resynthesized speech from the analyzed smooth time‐frequency surface and source parameters such as F0. This high‐quality resynthesis enables close investigations on naturalness deterioration as a function of feature modifications in the decomposed parameter domain; for example, detailed shape of a F0 trajectory, underlying parameters to determine F0 trajectory dynamics, group delay alignment of excitation pulses and aperiodicity/periodicity ratio of the excitation source and so on. One of potential advantages of this strategy is based on the fact that our perceptual function is highly nonlinear. The other source of advantage is virtually an independent parameter set which allows precise control of parameter deviations from the original analysis results. An overview of recent findings and modification demonstrations will be presented. [Work supported by CREST grant of Japanese Science and Technology Corporation.]

Towards more natural synthetic speech.

Abstract: This article reports the results of two experiments in which factors such as duration, amplitude and noise are manipulated, in order to achieve more natural utterances in synthetic speech. The participants were native speakers of English, instructed to judge the naturalness of the different versions of utterances generated throughout the manipulations. The results indicate that there are significant individual preferences, as well as classification principles other than conventional ones. There is evidence to believe that further research in this area will render positive results in the search for naturalness. The same principles could be applied to search for naturalness in the prosodic structure of the synthetic utterances. Advancement in this area will surely render improvements in Spoken Dialogue Systems.