<p>Vocabulary knowledge is a prerequisite to successful comprehension for native speakers and second language learners alike. Proper names, a peculiar and diverse group of lexical items, have long been the focus of discussion in general linguistics but have received practically no attention in second language vocabulary acquisition research. This study is the first attempt to assess whether proper names impact on second language learners' listening ability. First, I examine the question of how proper names can be adequately defined and discuss their semantic, structural, pragmatic and functional properties. I analyze proper names in light of the prototype theory and argue that personal, deity and pet names constitute the core of the proper name category. Names of places and enterprises occupy an intermediate position while names of events and artefacts are considered the least prototypical, i.e. peripheral members of the category. After identifying essential properties of prototypical proper names, I argue that in a spoken (as opposed to a written) text proper names cannot be considered automatically known items and place high demands on the listeners' cognitive resources. English as a second language (ESL) learners have to bring in a large amount of linguistic and encyclopaedic knowledge in order to cope with proper names in the flow of speech. I propose a 3-level model of such knowledge: recognition -> categorization -> referent properties. I then subject this model to empirical testing. The first experiment shows that among intermediate to advanced ESL learners the proper names recognition rate is around 60 percent. It is harder for ESL listeners to recognize proper names when the percentage of difficult common vocabulary in the text is high. The participants' proficiency level and the structure of a specific text were also found to affect the ability to recognize unfamiliar names. Well over a third of proper names are missed, which suggests that in real life listening, ESL learners mistake unknown common expressions for proper names and vice versa. In the second experiment, the participants' comprehension of a news story is tested under two conditions: Names Known (all proper names are familiar prior to listening) and Names Unknown (all proper names are unfamiliar). Results indicate that the presence of unfamiliar proper names hinders the intermediate to advanced proficiency learners' comprehension of a short news text as measured by immediate free recall and the ability to evaluate proper names related statements. The effect is local; it concerns comprehension of details, particularly those details that are associated with processing the proper names themselves. The Names Unknown group produced fewer details and more incorrect inferences in their recalls, scored significantly lower on the measure of proper names related comprehension, and selfreported a lower amount of comprehension. In contrast, the Names Known group produced more details and fewer incorrect inferences in their recalls, scored much higher on the measure of proper names related comprehension, and self-reported a greater degree of comprehension. The experiment also shows that participants in the Names Unknown treatment were not always able to ascertain from context what the referent of an unfamiliar proper name is, and in cases when they did, they could not extract as much information about the referent as the participants in the Names Known treatment had available. It is evidently unrealistic to expect ESL learners to determine what unfamiliar proper names refer to from context. On average, after 2-3 attempts at listening participants in the Names Unknown group were able to extract just over 40 percent of the information about the referents of unfamiliar proper names. Also participants' difficulty ratings of experimental tasks confirmed that the presence of unfamiliar proper names definitely makes the text seem harder to understand. The last experiment replicated the findings of the previous one on a larger sample. The Names Known group performed significantly better on open-ended questions and true-false-don't know statements. A substantial effect of unfamiliar proper names on the overall comprehension scores was found. Around 17 percent of the variance in the scores was accounted for by familiarity/lack of familiarity with proper names. The findings also provide some evidence in support of the claim that a name form that hints at the cognitive category its referent belongs to is less likely to adversely affect comprehension than a form that does not. Unfamiliar proper names contribute to raising the vocabulary threshold in second language listening, which should be taken into account by teachers, test-developers and other TESOL (teaching English to speakers of other languages) professionals.</p>
A Pronunciation Prior Assisted Vowel Reduction Detection Framework with Multi-Stream Attention Method