Downloads & Free Reading Options - Results

The Common Core assessments emphasize short essay constructed-response items over multiple-choice items because they are more precise measures of understanding. However, such items are too costly and time consuming to be used in national assessments unless a way to score them automatically can be found. Current automatic essay-scoring techniques are inappropriate for scoring the content of an essay because they either rely on grammatical measures of quality or machine learning techniques, neither of which identify statements of meaning (propositions) in the text. In this report, we introduce a novel technique for using domain-independent, deep natural language processing techniques to automatically extract meaning from student essays in the form of propositions and match the extracted propositions to the expected response. The empirical results indicate that our technique is able to accurately extract propositions from student short essays, reaching moderate agreement with human rater scores.

This collection of abstracts is part of a continuing series providing information on recent doctoral dissertations. The 20 titles deal with a variety of topics, including the following: semantic constructivity in children's comprehension; text-based inferences generated by children in reading written discourse; a review of reading comprehension research; the psycholinguistic processes of beginning readers; the relationship between discourse structures and critical reading; information processing and retrieval of adult college students; the effects of anaphoric references, typographically modified organizers, and oral prequestioning on reading comprehension; word analysis strategies for reading; the effect of question order on reading achievement; sentence characteristics that influence the comprehensibility of text; the effect of chunking on the recall of written discourse; types of information addition in the psycholinguistic process of reading; the relationship of knowledge of given information to reading comprehension; prose learning as a function of adjunct questions and individual differences; and the relationship of language cues used in silent reading to the semantic and syntactic cues revealed in oral reading miscues. (GW)

This collection of abstracts is part of a continuing series providing information on recent doctoral dissertations. The 20 titles deal with a variety of topics, including the following: semantic constructivity in children's comprehension; text-based inferences generated by children in reading written discourse; a review of reading comprehension research; the psycholinguistic processes of beginning readers; the relationship between discourse structures and critical reading; information processing and retrieval of adult college students; the effects of anaphoric references, typographically modified organizers, and oral prequestioning on reading comprehension; word analysis strategies for reading; the effect of question order on reading achievement; sentence characteristics that influence the comprehensibility of text; the effect of chunking on the recall of written discourse; types of information addition in the psycholinguistic process of reading; the relationship of knowledge of given information to reading comprehension; prose learning as a function of adjunct questions and individual differences; and the relationship of language cues used in silent reading to the semantic and syntactic cues revealed in oral reading miscues. (GW)

The current study has two aims. First, we aim to replicate Wong et al.’s (2022) finding that English-speakers are sensitive to hardness sound symbolism in Japanese onomatopoeias (mimetic words) for tactile senses. Second, we aim to extend the experimental design of Wong et al. (2022) with an additional variable of Japanese scripts (Katakana vs. Hiragana) to investigate whether English speakers are also sensitive to orthographic effects on sound symbolism in Japanese onomatopoeias. Sound symbolism is a cross-culturally (linguistically) observed phenomenon which refers to a non-arbitrary association between sound and meaning. It has classically been tested with forced choice tasks of name-shape correspondences, famously known as the maluma-takete or bouba-kiki effect (Köhler, 1929; Ramachandran & Hubbard, 2001). Fort and Schwartz (2022) examined this effect from the perspective of speech science and proposed that formant frequency differences may drive the cross-modal association between phonemes and visual properties. Their findings suggest that round objects produce sounds with a lower spectral balance and that are more temporally smooth, whereas spiky objects produce spectrally high and temporally less smooth sounds. This results in audiovisual regularities in our perception with bouba sounding more rounded as rounded materials would produce such a sound and kiki sounding more spiky as spiky materials would produce such a sound. Japanese is a language with many onomatopoeias, which are words imitating physical sounds or manners/states of objects and behaviours. These mimetic words are known to be sound symbolic, used sensitively in daily lives to describe or emphasize features of matter/objects, for example, loud laughter can be explained as “gera-gera” whilst quiet laughter can be described by “kusu-kusu” (Iwasaki et al., 2007). Wong et al. (2022) investigated sensitivity to Japanese sound symbolism in English speakers from USA/Singapore and native Japanese speakers. Participants were presented with 60 sound symbolic words, written in Japanese hiragana for native Japanese speakers and in the Roman alphabet for English speakers. These 60 words constitute of 24 conventional sound symbolic words selected from Japanese dictionary (Ono, 2007), 24 novel sound symbolism generated by a Genetic Algorithm (Shimizu et al., 2015; Doizaki et al., 2017) and selected as demonstrating strong congruence with hardness and softness by 12 Japanese speakers, and 12 non-sound-symbolic pseudowords created by combining four Japanese characters pseudo-randomly. Participants rated each item’s tactile sense of “hardness”, “warmness”, and “roughness” and the ratings were compared between these language populations. Both groups showed a similar pattern of ratings in tactile senses (especially on hardness), demonstrating the sound symbolism in Japanese mimetic words and sensitivity to this in an English speaking population. Wong et al. (2022) suggested that hardness sound symbolism may reflect the co-occurrence of multiple physical properties, for example, objects must be sufficiently hard to form spiky/angular shapes, i.e. a soft substance such as slime cannot create an angular shape. Summing up with the findings of Fort and Schwartz (2022), sound symbolism in roundedness and angularity may be driven by consistencies between audio and physical properties. This may then explain why hardness sound symbolism in Japanese words was perceived similarly by Japanese and English speakers. The current research aims to replicate these findings by adopting items and rating scales from Wong et al. (2022), but importantly adding auditory presentation (as opposed to transcription in the Roman alphabet) of the words. Orthography is one of the factors known to affect sound symbolism. Cuskley et al. (2017) explored how bouba/kiki judgements may be influenced by the matching based on curvedness/roundedness of letters (i.e. non-words with curved letters are matched to round shapes). Their findings indicated participants may judge the correspondences between shape and name by orthographic features of these names, as the letters that constitute bouba are more rounded than kiki and vice versa in angularity in the Roman script. This orthographic effect on sound-symbolism may be a more present phenomenon in Japanese, due to the fact that Japanese words can be written in two different (alpha)syllabaries: Hiragana and Katakana. There are general guidelines but not so strict rules as to which should be used for certain types of words. The writer can therefore flexibly decide which to use, for example with the purpose of emphasizing particular words or specific semantic features. Researchers have tested whether the orthography that is used influences Japanese speakers’ perception of words using various experimental designs. For example, using semantic differential scales, Iwahara et al. (2003) reported that when words describing objects were written in Hiragana this led to higher impressions of softness and roundedness, whereas Katakana led to an impression of higher hardness and angularity. This orthographic effect is supported by mathematical calculations of Japanese orthography handwriting, hiragana is indeed more rounded, and katakana is more angular (Komatsu et al., 2014). However, very little empirical research has been conducted on this phenomenon and it is also not known whether this is acquired through experience with Japanese language and scripts or is a more universal phenomena rooted in cross-modal visual-verbal processing of language. Therefore, using novel Japanese mimetic words from Wong et al. (2022), the current study will compare how Japanese mimetic words written in either hiragana or katakana will affect the rated “hardness” of the object that the word describes, in both native Japanese speakers and English speakers. The current experiment will present 24 Japanese novel mimetic words from Wong et al. (2022) to native Japanese speakers and English speakers with no prior experience with Japanese spoken or written forms. Half the written forms of these words will be presented in Japanese hiragana and half in katakana. Subjects will rate the hardness of a material described by each word in a rating scale of 0 (very soft) to 10 (very hard).

Includes bibliographies