Self-organizing semantic maps have been described before , and numerous methods exist to construct spatial representations of lexical knowledge (e.g., ). However, to our knowledge, this is the first objective approach to construct, based on available data, a simultaneous quantitative representation of synonymy and antonymy in a continuous metric space, whose dimensions have clearly identified general meanings. The low dimensionality of this semantic map indicates that, although thousands of distinct categories of meanings are conceivable, only very few apply to all contexts without a substantial domain-specific alteration of their semantic content. This limited number of general meanings is consistent with recent independent linguistic dimensional analyses and contrasts with the extensive lists of semantic categories represented in Roget's thesaurus and related or similar endeavors .
As Koolstra and Beentjes posit, "the findings indicate that young children can acquire elements of a foreign language through watching subtitled television programs" (p.58). Thus, one of the greatest contributions is that word recognition in the no subtitles condition was not superior to the subtitled condition, suggesting that the belief that reading subtitles might distract viewers from hearing English words may be just a belief. Another contribution regards implicit learning since Dutch fourth graders did show some knowledge of the English language though they had never been taught English before.
These findings lend support to the benefits of watching subtitled television programs at home for language learning purposes, especially considering that more English vocabulary was acquired by those who watched subtitled TV at home more often. In light of the diversity of the studies carried out in the last two decades focusing on language learning and the effects of subtitling or captioning, some research trends seem to emerge regarding 'the SLA/subtitling/captioning triangle'. The studies synthesized here suggest a focus on language domains and how the use of subtitled/captioned video materials tends to contribute to their development. Thus, overall focus has been placed on L2 reading comprehension, L2 listening comprehension and word recognition, and L2 vocabulary learning/acquisition. As expected based on the form of the energy function H, words with similar meanings have similar proportions on the principal components of the map, i.e. small angles between their vectors . In contrast, words with opposite meanings tend to have anti-parallel vectors.
In particular, synonyms and antonyms in MS English had median angles of 13° and 170°, respectively (means of 21° and 165°). Less than 3% of synonym pairs have angles greater than 90°, and less than 1% of antonyms have angles smaller than 90°. Upon checking, these exceptions revealed rare instances of questionable assignments in the source dictionary, which the map effectively "corrects". For example, opposite and harmonizing are listed as synonyms in MS English, but their angle on the map, 145°, suggests otherwise.
Although in most usage cases opposite and harmonizing would be considered antonyms the assignment as synonym in the source dictionary may still be appropriate in specific contexts . As an alternative example, hot and cool are typically antonyms (referring e.g. to weather or beverages), except when used idiomatically to describe an idea, a videogame, or a classmate. A total of 150 foreign language university learners participated in the study, who were second or fourth year learners of Spanish and Russian, and second year learners of Arabic and Chinese.
Participants watched documentaries about three animals - salmon, dolphins, and bears - and then were given vocabulary tests which had the same target words despite the different target languages after being pre-tested. Half of the vocabulary words were presented orally first, while the other half were presented in written form and vice-versa. Participants were asked to translate the target words into English, which had the same target words despite the different target languages.
Also, participants answered multiple-choice questions in English about the main points of the story to check for their overall comprehension. The study also aimed to check whether there would be vocabulary and phrase acquisition by the participants. With few exceptions, this was evident when the results for the Danish and French tests are scrutinized and considering that Danish is closer to Dutch. Results also suggest that when there are significant differences in the experimental conditions, it was always to the advantage of the condition with the foreign language in the soundtrack, possibly linked to "efficiency of information processing" (p. 242). This lends support to the need for the inclusion of specific features of individual differences in studies that focus on the effects of subtitled and captioned materials on language learning. Although the constructed semantic map reveals definitive semantics in each of its significant principal components, the vector associated with every word in the map should be interpreted as a "noisy" measure rather than an exact set of numerical values.
First, the positions of individual words on the map depend on the selection of available synonym-antonym links, which only constitute a small subset of all possible synonym-antonym links. Adding or deleting a link changes map coordinates of the corresponding words. Stated differently, any dictionary of synonyms and antonyms only provides sparse sampling of the onym graph.
Unlike most previous studies, our model was not tailored for a special practical purpose, but was constructed starting from basic principles. Our energy function was selected as the most parsimonious analytical expression corresponding to the concept of synonym and antonym vector alignment. The first term is the simplest analytical expression that attempts to align synonym vectors in parallel and antonym vectors in opposite directions. The last term is the lowest symmetric power term that is necessary to keep the distribution compact.
This conceptual framework significantly differs from the frameworks mentioned above, including LSA , LDA , Multidimensional Scaling , etc. Semantics of the principal dimensions of our map are reproduced across databases and languages. This is not a characteristic of any previously constructed vector semantic map in computational linguistics. The semantic characterization of the map principal components also enables a direct comparison across corpora, languages, and data types.
As mentioned earlier, the first three components, but not the fourth, demonstrate high consistency across independent corpora (MS English vs. WN English) and languages (cf. MS French, Table 1). To extend the comparison of principal semantic components to a quantitative measurement across additional corpora and languages, we also Google-translated the MS German and MS Spanish dictionaries into English. For the scope of this analysis, each corpus was limited to the set of words that overlapped with the MS English core dictionary. For example, the 15,783 MS English core words and the 20,477 WN English core words have 5926 terms in common.
For MS French, the overlap was 4704 English words, mapped onto from 19,944 French terms, representing approximately 30% of the MS French core dictionary. Many French words projected onto single English words, because word inflections are listed separately in the MS French thesaurus; the same occurred in German. We then extracted several correlation measures between the word coordinates from each of the separate semantic maps and the MS English map. Hence, one may argue that audio/video correlation might have had little or no impact on the participants' performance of word recognition. Also, participants reported that captioning helped them understand the story better, aiding them regarding their vocabulary/phrase acquisition, improving their language listening skill, and that it was an enjoyable way to learn English. Nevertheless, factors regarding their personal learning experiences showed no correlation with the listening comprehension test.
Verbatim captions and a multiple-choice based test were adopted to check for content comprehension. Questions required participants to provide informational paraphrases, basic deductions, or synonym identification of low-frequency words, after watching the video segments twice. The study involved 40 adult RFL learners - native speakers of English - who were divided into two groups and 70 adult ESL learners - who spoke nine different native languages - equally divided into test and control groups. Five participants were randomly chosen from each test group to engage in a 5-minute oral interview to retell any of the video segments. Such an increasing attention could be related to the fact that "subtitled television programs seem to provide a rich context for foreign language acquisition" (KOOLSTRA & BEENTJES, 1999, p. 51). The disciplinary field of Second Language Acquisition has witnessed an increasing interest in the investigation of the effects of subtitled and captioned audiovisual materials on domains of language learning/acquisition.
In this context, this paper seeks to provide a systematic review of recent studies related to language learning aspects aided by the instructional/experimental use of subtitled and captioned materials. The present paper draws on relevant literature in the field of SLA that interfaces with subtitling/captioning, while outlining their goals and main findings. This paper also aims to unveil which dimensions have merited scholar attention the most in the last two decades. Finally, some considerations are made regarding possible avenues for future research, taking into account the existing literature and underinvestigated issues.
A pool of terms likely related to a given word is constituted by all synonyms of synonyms or, more generally, "onyms of onyms" of that word. In particular, words which are onyms of onyms are usually in overlapping semantic domains, but not all words in overlapping domains are onyms of onyms. Having a synonym or antonym in common does not guarantee, but strongly indicates, that two words pertain to overlapping semantic domains. Thus, within the pool of onyms of onyms, one could expect angular information to be a powerful predictor of semantic content. To test this hypothesis, we sampled 20 words from MS English and WN English, and computed the cosines of their angle with each of their onyms of onyms. We then assigned the binary values of +1 and −1 to the onyms of onyms that were also reported as synonyms or antonyms, respectively.
The correlation between the cosines and binary values was statistically significant in all 40 cases . It is tempting to extrapolate these considerations and assume that proximity of two words in the map is sufficient to ensure a similarity of their meanings. Unrelated word pairs vastly outnumber synonyms (∼1500∶1) and antonyms (∼7400∶1). The majority of unrelated words pertains to separate semantic domains, and could not possibly be considered synonyms or antonyms.
Even the tail ends of their angle distribution constitute a disruptive confounder of the semantic relations. Stated differently, given a particular word, it is fair to assume that, among all related terms, synonyms will be concentrated in the neighborhood and antonyms in the antipodes. Nevertheless, unrelated words will still constitute the majority of terms even close to 0° and 180°. These unrelated words randomly end up in the proximity of a given term by virtue of their large number in the self-organizing reduction of the high number of initial dimensions into the low-dimensional principal component space.
Our low-dimensional map complements those local approaches by observing global semantic properties. Here, the distance between locations selectively measures the aspects of the dissimilarity broadly applicable to any context, without distinguishing between domain-specific semantic flavors. Children vocabulary acquisition in a foreign language through watching subtitled television programs at home. The participants in the captioned TV group consistently achieved higher mean scores than the other groups on all word knowledge tests, though these differences were not always statistically significant. Comprehensible input, facilitated in a bimodal input form (audio + image), and the helpfulness of the context seem to have played a key role in the participant's vocabulary acquisition.
As a result of these two limitations, namely sparse sampling and approximation of meanings with words, individual word coordinates are subject to considerable noise, the relative amplitude of which can be roughly estimated as 10–20%. Nevertheless, the map is robust with respect to the assignments of synonyms and antonyms, and their connotation, from sets of related words. In particular, within all onyms of onyms, constituting a pool of terms likely related to a given word, dot product is a powerful predictor of semantic content . Moreover, when global map characteristics are derived from all word coordinates, as in cross-corpus map correlations the noise effectively averages out. This means that the map can be used as a precise semantic scale, even if individual words cannot.
The second component of the map similarly orders terms based on a connotation of "calming-exciting" or "easy-difficult" (vertical axis in Figure 1A–C). Both the sign and the relative value of this coordinate are again consistent semantic predictors, as in the examples of relax (−1.55 in the MS map, −1.05 in WN), troubling (0.62, 0.95), and excite (0.99, 1.16). Since principal components are by construction orthogonal on the map, the values of word coordinates in these first two dimensions (PC#1 and PC#2) are mutually independent. More generally, while the positions of words in the maximum spread projection are highly consistent among MS English, WordNet, and the map derived from MS French thesaurus (Figure 1A–C), they bear no implication on the values of subsequent components . Here xi is the 26-dimensional vector representing the ith word in the configuration x.
The Wij entries of the symmetric relation matrix equal +1 for pairs of synonyms, −1 for pairs of antonyms, and zero for all non-onym pairs. Intuitively, maximizing the first sum moves synonyms towards the same hemispaces, while minimizing the second tends to align antonym pairs on opposite sides of the origin, reflecting their semantic relations. The fourth-power norm provides a soft limit to the absolute distance from the center. More specifically, the first term of the equation is the simplest analytical expression that captures the intent of aligning synonym vectors in parallel and antonym vectors in opposite directions.
This general approach and specific selection were empirically validated by their successful reconstruction of a map whose meaning was known a priori, that of color space, as illustrated at the end of the Results section. The vast array of topics covered suggests how prolific the effects of subtitling and captioning can be on language learning/acquisition. Vocabulary acquisition scores were higher in the subtitled condition and scores in the no subtitles condition were higher than in the control group. The analysis also yielded a main effect of grade, with sixth graders outperforming fourth graders. Concerning word recognition, more English words were recognized after participants watched the subtitled television program in comparison to the no subtitles condition, and sixth graders outperformed fourth graders in this test as well. There was indeed a correlation between high frequency of watching subtitled programs at home often and higher English vocabulary scores.
The captioned TV group outscored the reading text group for all three units on word recognition, though they were not statistically significant for Units 1 or 3. Regarding the results of the sentence anomaly tests, differences among all four groups indicated a similar trend, favoring captioning. As for unit tests, results revealed that the captioned TV group outperformed the reading text group, but when comparing the captioned TV group and the traditional TV group, results were only statistically significant for Unit 2. These results indicate that specific features inherent to the videos in each unit must have played a role in the students' comprehension of the science topics. As they speculate, "the visual representation of words in video form is an important contributor to students' increased word knowledge" (p. 102). A potential practical application of the described semantic map consists of specifying the connotation as well as the general meaning of words.
An illustration of considering connotation is provided in Figure 6, where onyms of onyms of two words are plotted in the plane of the first two principal components. In general, terms are located in the proper octant according to the connotation of their meaning ("good", "good/exciting", "exciting", "bad/exciting", "bad", "bad/calming", "calming", "good/calming"). For instance, the term control can be substituted with a "good" connotation by organize, or with a "bad" connotation as curb. Likewise, delicate can connote a "calming" semantic as soft or an "exciting" semantic as personal. Words of natural language along with idioms and phrases are used in speech and writing to communicate conscious experiences, such as thoughts, feelings, and intentions.
Each meaningful word, considered without any context, is characterized by a set of semantic connotations . These connotations are a product of, and correlate with experiences communicated with the use of the word. Stated differently, communicated word semantics are behavioral correlates of experienced semantics.
Therefore, the scientific characterization of word semantics can shed light on semantics of human experiences. In particular, if word meaning can be measured based on a metric system, the same metric system might be useful to measure the meaning of experiences. Thus, a precise metric system for the semantics of words could be a key in developing empirical science of the human mind .
This paper set out to provide a systematic review of recent studies addressing language learning aspects aided by the instructional and experimental use of subtitled and captioned video materials. This paper also attempted to unveil which topics have merited scholar attention in the past years, as well as pointing out underinvestigated issues in this particular niche. The study addressed the effects of single modality - either sound or text - and bimodal input - sound and text - presentation on word learning, with an explicit focus on word learning.
Measures involved enhancements in spoken word recognition efficiency and recognition memory related to word retention . Results are suggestive of the positive effects of captions on comprehension for the striking number of correct answers reveals that the presence of captions significantly increased the amount of comprehensible input to the foreign language learners. Captions facilitated learners' comprehension of the video they watched. Nonetheless, when comparing the performance of the groups, it is clear that the presence of captions had a more significantly positive effect on the Russian group when they watched for the second time with captions. Yet, the number of correct responses of the ESL non-captioned group was still higher than the number of correct responses by the Russian group when watching for the second time. In this context, this paper seeks to provide a brief, systematic review of recent studies3 that address language learning aspects aided by the instructional/experimental use of subtitled and captioned materials.
This state-of-the-art paper synthesizes the relevant and groundbreaking literature in the field of SLA that interfaces with subtitling and captioning, while outlining their goals and main findings. It also aims to unveil which topics have merited scholar attention the most in the past years. Finally, some considerations are made as to possible avenues for future research, taking into account the existing literature and underinvestigated issues.
The semantic similarity of our map with ANEW in the first two dimensions was quantitatively confirmed by canonical correlation analysis, based on the map locations of words that are common for the two maps. The map constructed in the present study contains more dimensions and more words, including words that do not belong to affective stimuli. Most importantly, this map differs qualitatively from previous data as it was not constructed based on given semantic dimensions. Instead, semantics of our map dimensions are emergent and defined by the locations of all words together.
