Probabilistic Reduction Research Articles

Probabilistic reduction and constructionalization: a usage-based diachronic account of the diffusion and conventionalization of the Spanish la de <noun> que construction

Probabilistic reduction and constructionalization: a usage-based diachronic account of the diffusion and conventionalization of the Spanish <i>la de</i> &lt;noun&gt; <i>que</i> construction

Knowledge from Probability

We give a probabilistic analysis of inductive knowledge and belief and explore its predictions concerning knowledge about the future, about laws of nature, and about the values of inexactly measured quantities. The analysis combines a theory of knowledge and belief formulated in terms of relations of comparative normality with a probabilistic reduction of those relations. It predicts that only highly probable propositions are believed, and that many widely held principles of belief-revision fail.

Probabilistic reduction and mental accumulation in Japanese: Frequency, contextual predictability, and average predictability

A large number of studies have demonstrated that a linguistic unit with higher frequency and that with higher contextual predictability are realized with reduced speech signals. This effect is known in previous literature as probabilistic reduction. A few studies also demonstrate that these contextual effects may be memorized; that is, average predictability may also influence phonetic redundancy in speech signals. This study aims to investigate the extent to which probabilistic reduction effects extend to morpheme duration in Japanese. It is demonstrated that the production of a morpheme in Japanese is influenced by morpheme frequency, contextual morpheme predictability, and average morpheme predictability. These findings are interpreted to mean that speech signals are influenced by ease of lexical access or they are modified to maximize communicative ease, and the productions are accumulated in the mind of a speaker.

Prosody leaks into the memories of words

The average predictability (aka informativity) of a word in context has been shown to condition word duration (Seyfarth, 2014). All else being equal, words that tend to occur in more predictable environments are shorter than words that tend to occur in less predictable environments. One account of the informativity effect on duration is that the acoustic details of probabilistic reduction are stored as part of a word's mental representation. Other research has argued that predictability effects are tied to prosodic structure in integral ways. With the aim of assessing a potential prosodic basis for informativity effects in speech production, this study extends past work in two directions; it investigated informativity effects in another large language, Mandarin Chinese, and broadened the study beyond word duration to additional acoustic dimensions, pitch and intensity, known to index prosodic prominence. The acoustic information of content words was extracted from a large telephone conversation speech corpus with over 400,000 tokens and 6000 word types spoken by 1655 individuals and analyzed for the effect of informativity using frequency statistics estimated from a 431 million word subtitle corpus. Results indicated that words with low informativity have shorter durations, replicating the effect found in English. In addition, informativity had significant effects on maximum pitch and intensity, two phonetic dimensions related to prosodic prominence. Extending this interpretation, these results suggest that predictability is closely linked to prosodic prominence, and that the lexical representation of a word includes phonetic details associated with its average prosodic prominence in discourse. In other words, the lexicon absorbs prosodic influences on speech production.

Predictability modulates pronunciation variants through speech planning effects: A case study on coronal stop realizations

Predictability has been shown to be associated with many dimensions of variation in speech, including durational variation and variable omission of segments. However, the mechanism or mechanisms that underlie these effects are still unclear. This paper presents data on a new aspect of predictability in speech, namely how it affects allophonic variation. We examine two coronal stop allophones in English, flap and glottal stop, and find that their relationship with predictability is quite different from what is expected under current theories of probabilistic reduction in speech. Flapping is more likely when the word that follows is more predictable, but is not influenced by the frequency of the word itself, while glottal stops are more likely in words that are less predictable. We propose that the crucial distinction between these two allophones is how they are conditioned by phonological context. This, we argue, interacts with online speech planning processes and gives rise to variability for context-dependent allophones. This hypothesis offers a specific, testable mechanism for certain predictability effects, and has the potential to extend to other factors that contribute to variability in speech.

Entanglement accessibility measures for the quantum Internet

We define metrics and measures to characterize the ratio of accessible quantum entanglement for complex network failures in the quantum Internet. A complex network failure models a situation in the quantum Internet in which a set of quantum nodes and a set of entangled connections become unavailable. A complex failure can cover a quantum memory failure, a physical link failure, an eavesdropping activity, or any other random physical failure scenarios. Here, we define the terms such as entanglement accessibility ratio, cumulative probability of entanglement accessibility ratio, probabilistic reduction of entanglement accessibility ratio, domain entanglement accessibility ratio, and occurrence coefficient. The proposed methods can be applied to an arbitrary topology quantum network to extract relevant statistics and to handle the quantum network failure scenarios in the quantum Internet.

Probabilistic reduction in relation to social message predictability

AbstractLanguage is a system of message transmission, which conveys a variety of messages including both lexical messages and social messages. It has been demonstrated that lexical messages are realized with phonetically reduced signals, when they are contextually predictable. For example, a word may be produced with shorter duration, when it is more predictable given a context such as a preceding word and a following word. This message-oriented reduction can be encapsulated by positing that a speaker is required to balance two biases: a bias for maximizing the accuracy of message transmission and a bias for maximizing the efficiency of message transmission. This raises a question: Does a speaker balance the two biases in relation to social messages? The aim of this study is to address this question, and advance our understanding of the message-oriented probabilistic reduction. We will explore the social-message-predictability effects by examining the phonetic redundancy in a variant in New Zealand English loanword phonology, a tap sound [ɾ]. It is demonstrated that the duration of this rhotic variant is affected by the social message predictability given a loanword.

Confluence and convergence modulo equivalence in probabilistically terminating reduction systems

Convergence of an abstract reduction system is the property that the possible derivations from a given initial state all end in the same final state. Relaxing this by “modulo equivalence” means that these final states need not be identical, only equivalent wrt. a specified equivalence relation.We generalize this notion for probabilistic abstract reduction systems, naming it almost-sure convergence modulo equivalence, such that the final states are reached with probability 1. We relate it to the well-studied properties of almost-sure termination and confluence/convergence of probabilistic and non-probabilistic systems. In addition, we provide a transformational approach for proving – or disproving – almost-sure convergence modulo equivalence of given systems.

Patterns of probabilistic segment deletion/reduction in English and Japanese

Abstract Probabilistic phonetic reduction is widely attested in a variety of languages, acoustic domains, and interpretations of predictability. Less well-studied is the categorical effect of probabilistic segment deletion, which in principle is subject to similar pressures. This paper presents the results of an exploratory study into patterns of segment deletion in corpora of spontaneous speech in English and Japanese. Analysis at the word level reveals that words with more phonemes and higher-frequency words tend to have more of their segments deleted. Analysis at the phoneme level reveals that high-probability phonemes are more likely to be deleted than low-probability phonemes. For Japanese only, this analysis also shows effects of word length, frequency, and neighborhood density on deletion probability. Taken together, these results suggest that several large-scale patterns of probabilistic segment deletion mirror the processes of phonetic reduction and apply to both languages. Some patterns, though, appear to be language-specific, and it is not clear to what extent languages can and do differ in this regard. These findings are discussed in terms of our understanding of the universality of proposed predictability effects, and in terms of probabilistic reduction more broadly.

Contextual predictability influences word and morpheme duration in a morphologically complex language (Kaqchikel Mayan).

The probability is one of the many factors which influence phonetic variation. Contextual probability, which describes how predictable a linguistic unit is in some local environment, has been consistently shown to modulate the phonetic salience of words and other linguistic units in speech production (the probabilistic reduction effect). In this paper the question of whether the probabilistic reduction effect, as previously observed for majority languages like English, is also found in a language (Kaqchikel Mayan) which has relatively rich morphology is explored. Specifically, whether the contextual predictability of words and morphemes influences their phonetic duration in Kaqchikel is examined. It is found that the contextual predictability of a word has a significant effect on its duration. The effect is manifested differently for lexical words and function words. It is also found that the contextual predictability of certain prefixes in Kaqchikel affects their duration, showing that contextual predictability may drive reduction effects at multiple levels of structure. While the findings are broadly consistent with many previous studies (primarily on English), some of the details of the results are different. These differences highlight the importance of examining the probabilistic reduction effect in languages beyond the majority, Indo-European languages most commonly investigated in experimental and corpus linguistics.

Effects of Word Frequency and Transitional Probability on Word Reading Durations of Younger and Older Speakers

High-frequency units are usually processed faster than low-frequency units in language comprehension and language production. Frequency effects have been shown for words as well as word combinations. Word co-occurrence effects can be operationalized in terms of transitional probability (TP). TPs reflect how probable a word is, conditioned by its right or left neighbouring word. This corpus study investigates whether three different age groups-younger children (8-12 years), adolescents (12-18 years) and older (62-95 years) Dutch speakers-show frequency and TP context effects on spoken word durations in reading aloud, and whether age groups differ in the size of these effects. Results show consistent effects of TP on word durations for all age groups. Thus, TP seems to influence the processing of words in context, beyond the well-established effect of word frequency, across the entire age range. However, the study also indicates that age groups differ in the size of TP effects, with older adults having smaller TP effects than adolescent readers. Our results show that probabilistic reduction effects in reading aloud may at least partly stem from contextual facilitation that leads to faster reading times in skilled readers, as well as in young language learners.

Quantifying travel time variability at a single bottleneck based on stochastic capacity and demand distributions

ABSTRACTTravel time reliability, an essential factor in traveler route and departure time decisions, serves as an important quality of service measure for dynamic transportation systems. This article investigates a fundamental problem of quantifying travel time variability from its root sources: stochastic capacity and demand variations that follow commonly used log-normal distributions. A volume-to-capacity ratio-based travel time function and a point queue model are used to demonstrate how day-to-day travel time variability can be explained from the underlying demand and capacity variations. One important finding is that closed-form solutions can be derived to formulate travel time variations as a function of random demand/capacity distributions, but there are certain cases in which a closed-form expression does not exist and numerical approximation methods are required. This article also uses probabilistic capacity reduction information to estimate time-dependent travel time variability distributions under conditions of non-recurring traffic congestion. The proposed models provide theoretically rigorous and practically useful tools for understanding the causes of travel time unreliability and evaluating the system-wide benefit of reducing demand and capacity variability.

Amplification uncertainty relation for probabilistic amplifiers

Traditionally, quantum amplification limit refers to the property of inevitable noise addition on canonical variables when the field amplitude of an unknown state is linearly transformed through a quantum channel. Recent theoretical studies have determined amplification limits for cases of probabilistic quantum channels or general quantum operations by specifying a set of input states or a state ensemble. However, it remains open how much excess noise on canonical variables is unavoidable and whether there exists a fundamental trade-off relation between the canonical pair in a general amplification process. In this paper we present an uncertainty-product form of amplification limits for general quantum operations by assuming an input ensemble of Gaussian distributed coherent states. It can be derived as a straightforward consequence of canonical uncertainty relations and retrieves basic properties of the traditional amplification limit. In addition, our amplification limit turns out to give a physical limitation on probabilistic reduction of an Einstein-Podolsky-Rosen uncertainty. In this regard, we find a condition that probabilistic amplifiers can be regarded as local filtering operations to distill entanglement. This condition establishes a clear benchmark to verify an advantage of non-Gaussian operations beyond Gaussian operations with a feasible input set of coherent states and standard homodyne measurements.

Probabilistic reduction and probabilistic enhancement

Research on probabilistic pronunciation variation has generally focused on contextual probability, or the probability of using a linguistic unit (segment, syllable, word, etc.) in the context of a particular utterance. Overwhelmingly, this research has found that higher contextual probability leads to phonetic reduction. Less attention, however, has been given to paradigmatic probability, or the probability of using a particular linguistic form from a paradigm of related forms. The research that has addressed this type of probability has found inconsistent results: Sometimes higher paradigmatic probability leads to phonetic enhancement, and sometimes to phonetic reduction. In this paper I present the results of an experiment exploring the effects of both types of probability simultaneously on the pronunciation of agreement suffixes on English verbs. I find that (i) singular verb suffixes and stems are phonetically reduced when singular agreement is contextually probable; (ii) the nature of the reduction is modulated by verb frequency, consistent with dual-route models of lexical retrieval; and (iii) suffixes are phonetically enhanced when they are paradigmatically probable. I conclude by discussing how the patterns observed in this study shed light on the previous contradictory findings regarding the effects of paradigmatic probability on pronunciation.

Word informativity influences acoustic duration: Effects of contextual predictability on lexical representation

Language-users reduce words in predictable contexts. Previous research indicates that reduction may be stored in lexical representation if a word is often reduced. Because representation influences production regardless of context, production should be biased by how often each word has been reduced in the speaker’s prior experience. This study investigates whether speakers have a context-independent bias to reduce low-informativity words, which are usually predictable and therefore usually reduced. Content word durations were extracted from the Buckeye and Switchboard speech corpora, and analyzed for probabilistic reduction effects using a language model based on spontaneous speech in the Fisher corpus. The analysis supported the hypothesis: low-informativity words have shorter durations, even when the effects of local contextual predictability, frequency, speech rate, and several other variables are controlled for. Additional models that compared word types against only other words of the same segmental length further supported this conclusion. Words that usually appear in predictable contexts are reduced in all contexts, even those in which they are unpredictable. The result supports representational models in which reduction is stored, and where sufficiently frequent reduction biases later production. The finding provides new evidence that probabilistic reduction interacts with lexical representation.

삼원 촉매 담체의 확률론적 열피로 성능 평가

A three-way catalyst substrate for domestic passenger car satisfies the design criteria for exhaust gas exchange and pressure drop but does not have satisfactory thermal fatigue performance. Prefracture faults in this three-way catalyst substrate has often been discovered in vehicle repair or vehicle inspection facilities. This paper presents a thermal fatigue performance estimation method for a three-way catalyst substrate using a probabilistic strength reduction factor model. This method is superior to the thermal fatigue performance estimation method for a three-way catalyst substrate that uses a deterministic strength model.

The probabilistic reduction approach to specifying multinomial logistic regression models in health outcomes research

The paper provides a novel application of the probabilistic reduction (PR) approach to the analysis of multi-categorical outcomes. The PR approach, which systematically takes account of heterogeneity and functional form concerns, can improve the specification of binary regression models. However, its utility for systematically enriching the specification of and inference from models of multi-categorical outcomes has not been examined, while multinomial logistic regression models are commonly used for inference and, increasingly, prediction. Following a theoretical derivation of the PR-based multinomial logistic model (MLM), we compare functional specification and marginal effects from a traditional specification and a PR-based specification in a model of post-stroke hospital discharge disposition and find that the traditional MLM is misspecified. Results suggest that the impact on the reliability of substantive inferences from a misspecified model may be significant, even when model fit statistics do not suggest a strong lack of fit compared with a properly specified model using the PR approach. We identify situations under which a PR-based MLM specification can be advantageous to the applied researcher.

Do speakers articulate over-described modifiers differently from modifiers that are required by context? Implications for models of reference production

Studies have shown that speakers often include unnecessary modifiers when producing referential expressions, which is contrary to the Maxim of Quantity. In this study, we examined the production of referring expressions (e.g. the red triangle) that contained an over-described (or redundant) pre-nominal adjective modifier. These expressions were compared to similar expressions that were uttered in a context that made the modifier necessary for unique referent identification. Our hypothesis was that speakers articulate over-described modifiers differently from those used to distinguish contrasting objects. Results showed that over-described modifiers were significantly shorter in duration than modifiers used to distinguish two objects. Conclusions focus on how these acoustic differences can be modelled by Natural Language Generation algorithms, such as the Incremental Algorithm, in combination with probabilistic prosodic reduction.

Phonological encoding and articulatory duration in spontaneous speech

Many studies have found that word duration is correlated with a word’s contextual predictability in conversational speech. Lindblom (1990)’s Hypo/Hyperarticulation Theory, Jurafsky et al. (2001)’s Probabilistic Reduction Hypothesis, and Aylett and Turk (2006)’s Smooth Signal Redundancy Hypothesis all suggest that such differences in duration are due to processes occurring primarily at the lexical level. The present study, however, suggests that these differences may be attributable to processes occurring at the phonological level. In this study, mixed modeling is used to examine the voice onset time and rime duration of monosyllabic /p t k/ words in spontaneous, connected speech (the Buckeye Corpus; Pitt et al., 2007). Higher contextual predictability given the previous word is found to be associated with shorter VOT, while higher contextual predictability given the following word is associated with shorter rime duration. VOT also varies according to the number and type of a word’s phonological neighbors; words with more neighbors overlapping in the rime have significantly longer VOT, while words with more neighbors overlapping in the initial CV have significantly shorter VOT. These results motivate a model of speech production that assumes both lexical-phonological feedback and positional encoding of segments (e.g., Sevald and Dell, 1994).

A cautionary note on automated statistical downscaling methods for climate change

The urge for higher resolution climate change scenarios has been widely recognized, particularly for conducting impact assessment studies. Statistical downscaling methods have shown to be very convenient for this task, mainly because of their lower computational requirements in comparison with nested limited-area regional models or very high resolution Atmosphere–ocean General Circulation Models. Nevertheless, although some of the limitations of statistical downscaling methods are widely known and have been discussed in the literature, in this paper it is argued that the current approach for statistical downscaling does not guard against misspecified statistical models and that the occurrence of spurious results is likely if the assumptions of the underlying probabilistic model are not satisfied. In this case, the physics included in climate change scenarios obtained by general circulation models, could be replaced by spatial patterns and magnitudes produced by statistically inadequate models. Illustrative examples are provided for monthly temperature for a region encompassing Mexico and part of the United States. It is found that the assumptions of the probabilistic models do not hold for about 70 % of the gridpoints, parameter instability and temporal dependence being the most common problems. As our examples reveal, automated statistical downscaling “black-box” models are to be considered as highly prone to produce misleading results. It is shown that the Probabilistic Reduction approach can be incorporated as a complete and internally consistent framework for securing the statistical adequacy of the downscaling models and for guiding the respecification process, in a way that prevents the lack of empirical validity that affects current methods.