DISCUSSION AND CONCLUSION
讨论与结论

Regarding the primary research question, when learners were asked to sum up the content of the TED Talk before watching it a second time, they gained knowledge of 3 (16.67%) and 8 (44.44%) out of the18target words on average according to the strict scoring method and
关于主要研究问题，当学习者被要求在观看第二次 TED 演讲之前总结其内容时，根据严格的评分方法，他们平均对 18 个目标词汇中的 3 个（16.67%）和 8 个（44.44%）获得了知识。

the scoring method that gave credit for partial knowledge, respectively. Although arguably modest, this vocabulary gain nonetheless surpassed the gain attested in the comparison condition, where learners also watched the TED Talk twice but without insertion of an oral output task. In the latter condition, the students learned on average two(11.11%) and four (22.22%)words according to the two scoring proto- cols. The students in the summarycondition also obtained signifi- cantly better scores on the text comprehension test than the comparison group. It is conceivable that the announced summary task helped the learners process the input material with a clearer purpose in mind than only the announcement that content questions would follow. It is also possible that the actual retelling helped them consoli- date the information gathered from the first viewing(Wittrock, 1991). In addition, if the learners felt their summary was incomplete or per- haps not fully accurate, this may have prompted them to seek the miss- ing information in the video during the second viewing.
给分方法为认可部分知识，分别。尽管可能被认为较为保守，但词汇量的增加仍然超过了比较条件下的增加，其中学习者也观看了两次 TED 演讲，但没有插入口头输出任务。在后者条件下，根据两种评分协议，学生平均学习了两个（11.11%）和四个（22.22%）单词。在总结条件下的学生在文本理解测试中的得分也明显优于比较组。可以想象，宣布的总结任务帮助学习者在比仅宣布后续内容问题时更有明确目的地处理输入材料。实际上，重述也可能帮助他们巩固了从第一次观看中收集的信息（Wittrock, 1991）。此外，如果学习者觉得他们的总结不完整或可能不够准确，这可能促使他们在第二次观看时寻找缺失的信息。

As far as the secondary research questions are concerned, the sum- mary task certainly created a near-immediate opportunity for the learn- ers to try using newly met words, and--unlike what was attested in Yang et al. (2017), where the chosen output task was more loosely con- nected to the input text--the students made use of this opportunity. Crucially, when learners used target words in their summaries, they also managed to recall the meaning of these words at least up to 2 weeks later. This may attest to the usefulness of retrieval and of the generative use of new words, two desirable features of vocabulary learning activities according to Nation and Webb's (2011) technique feature analysis. This finding is also in keeping with Laufer and Hul- stijn's (2001) involvement hypothesis because the active use of a newly met word entails a certain evaluation on the part of the learner of its accuracy and appropriateness in the given context. When the word is subsequently reencountered in the model text, further evaluation of one's own use of the word is likely to happen.
关于次级研究问题，总结任务确实为学习者提供了立即尝试使用新遇到的单词的机会，与杨等人（2017 年）中选择的输出任务与输入文本联系不太紧密不同，学生们利用了这个机会。至关重要的是，当学习者在总结中使用目标单词时，他们也能够至少在两周后回忆起这些单词的含义。这可能证明了检索和新词生成性使用在词汇学习活动中的有用性，根据纳特和韦布（2011 年）的技术特征分析。这一发现也与劳弗和胡尔廷（2001 年）的参与假设相一致，因为新遇到的单词的主动使用意味着学习者对其在给定上下文中的准确性和适当性的评估。当随后在模型文本中遇到该单词时，对自身使用该单词的进一步评估很可能会发生。

Moreover, even where learners tried but failed to incorporate target words in their oral summaries, the output task seems to have been indirectly conducive to learning. That, at least, is suggested by the many episodes of hesitations suggesting efforts of retrieval that were followed by successful recall in the posttests. The latter finding is in accordance with the output hypothesis and again also illustrates the usefulness of revisiting the input text that a given output task is based on so as to give learners the opportunity to mine that input for ele- ments they felt lacking in their expressive resources.
此外，即使在学习者尝试但未能在口头总结中融入目标词汇的情况下，输出任务似乎间接促进了学习。至少，许多犹豫的片段表明了检索尝试，随后在后续测试中成功回忆，这表明了这一点。这一发现与输出假设一致，再次证明了回顾基于输出任务的输入文本的有用性，以便给学习者机会从其表达资源中挖掘缺失的元素。

It is worth emphasizing that the study reported here was a class- room-based experiment. It is unlikely that learners will often engage in content-retelling after they have watched a video outside theclassroom-unless they really feel the desire to tell someone who has not watched it. As a classroom activity (or as an activity incorporated in an online interactive course), it is relatively easy to implement, how- ever. Although (for the sake of experimental control)the participants in the present study all watched the same video and recorded their summaries individually in the form of monologues, in more genuinely communicative practice students could be asked to watch different videos on a related topic and then be paired up to exchange informa- tion about them.
值得强调的是，这里报告的研究是一个基于课堂的实验。不太可能在学生在课后观看视频后经常进行内容复述，除非他们真的渴望告诉那些没有看过视频的人。作为课堂活动（或作为在线互动课程中的一项活动），它相对容易实施。尽管（为了实验控制），本研究中的所有参与者都观看了相同的视频，并以独白的形式各自记录了他们的总结，但在更真实的交际实践中，学生可以被要求观看与主题相关的不同视频，然后配对交换关于这些视频的信息。

Although the chosen output task enhancedvocabulary learning from the TED Talk, it needs to be acknowledged that theoutcome in terms of the number of new words actually learned still remains mod- est. Many factors may help account for this. First, the study material in this experiment was authentic in the sense that it was produced with- out L2 learners in mind. Such characteristics of authentic materials as natural and fast speech must be challenging to learners. It is worth noting in this regard that much previous research on incidental vocab- ulary acquisition used either special-purpose instructional materials, like graded readers (e.g., Brown et al., 2008; Horst et al., 1998) or materials developed specifically for research purposes (e.g., van Zee- land& Schmitt, 2013; Vidal, 2011). For instance, becausethe video in this experiment was not modified, word repetition in the input text was not manipulated. Consequently, all but l of the 18 target words(bioluminescence) occurred only once or twice in the text, which is much lower than the 10+ or 20+times suggested by van Zeeland and Schmitt (2013, p.621) or Brown et al.(2008, p. 153) as necessary for incidental vocabulary acquisition to occur. Second, although learners with a similar profile as the actual participants had suggested that the selected TED Talks video was suitable for use in this study (see the Input Material and Target Words section above),the comprehension test scores indicatedthat its content was far from straightforward. At the same time, the TED Talk chosen for the experiment illustrates some of the affordances of the genre for L2vocabulary acquisition more generally. For example, one of the words that was relatively well remembered was bioluminescence. In the TED Talk, the meaning of this word was clearly denoted via a verbal explanation(“So it is a little appreciated fact that most of the animals in our ocean make light. I have spent most of my career studying this phenomenon called biolu- minescence.") and a series of corresponding footage. This item might also have stood a better chance of learning as it was essential for text comprehension and occurred six times in the given passage. Another example was the word lure in the phrase optical lure. In the video, the written form of this phrase was put on the screen and its meaning was both visually and verbally demonstrated, as in “This is an optical lurethat I have used. We call it the electronic jellyfish. It is just 16 blue LEDs that we can program to do different types of [light] display.” It would be interesting to be able to compare the learning rate observed in the present study to the learning rates reported in studies such as those listed in Table 1, where the participants engaged in more deliberate postreading work on vocabulary. Such comparisons should in any case be interpreted very cautiously because of differ- ences in materials, procedures, participants, and so forth, but they are particularly difficult here owing to the diversity of test formats used. One study in Table 1, Hulstijn and Laufer(2001),used a meaning- recall test like we did and also gave credit for partially correct responses. The task that led to the best learning gains in two trials of the same experiment (conducted with different student populations) was writing a text incorporating 10 target words from a glossed reading text. Total time spent on the reading and writing tasks was 70-80 minutes. The delayed posttest scores were 2.6 in one trial and 3.7 in the other. The success rate attested in theexperiment we have reported here does not compare unfavorably to this considering that it was obtained through classroom procedures that took only about 40 minutes. Lu (2013) also included a meaning-recall test (but responses were scored only dichotomously). The students read a story containing nine glossed target words, and then performed tasks with a focus on these nine words, such as summary writing incorporating the words(which took close to 40 minutes in addition to the time spent on reading the story) and completing three gapped sentences for each word (which took close to 20minutes in addition to the time spent on the story). On average, the students in these two conditions recalled the meaning of two of the nine target words in the delayed posttest. Again, the learning we have observed in our experiment, without input modification and without instruction to engage stu- dents in deliberate word learning, seems not to compare unfavorably to this.
尽管选择的输出任务增强了从 TED 演讲学习词汇，但仍需要承认，实际学到的新词汇数量的成果仍然相对有限。许多因素可能有助于解释这一点。首先，实验中的学习材料是真实的，因为它没有考虑到第二语言学习者的需求。这类真实材料的自然和快速口语特性对学习者来说可能是具有挑战性的。值得注意的是，关于偶然词汇习得的大量先前研究使用了专门设计的教学材料，如分级读物（例如，Brown 等，2008 年；Horst 等，1998 年）或为研究目的专门开发的材料（例如，van Zeland& Schmitt，2013 年；Vidal，2011 年）。例如，由于实验中的视频没有被修改，输入文本中的词汇重复次数没有被操纵。因此，除了“生物发光”这个词外，18 个目标词汇中的所有词汇在文本中只出现了一次或两次，这远低于 van Zeeland 和 Schmitt（2013 年，第 621 页）或 Brown 等（2008 年，第 153) 为了偶然词汇学习的发生所必需。其次，尽管与实际参与者具有相似特征的学习者曾建议所选的 TED 演讲视频适合用于这项研究（请参阅上方的输入材料和目标词汇部分），但理解测试分数表明其内容远非简单。同时，为实验所选择的 TED 演讲展示了该体裁在第二语言词汇学习方面的一些优势。例如，一个相对容易记住的词汇是生物荧光。在 TED 演讲中，通过口头解释（“有趣的是，我们海洋中的大多数动物都会发光。我大部分职业生涯都在研究这种名为生物荧光的现象。”）和一系列相应的视频片段，这个词汇的意义被明确指出了。这个项目可能也有更好的学习机会，因为它对于文本理解至关重要，并在给定的段落中出现了六次。另一个例子是短语光学诱饵中的单词诱饵。 在视频中，这句话的书写形式被放在屏幕上，并通过视觉和口头的方式展示了其含义，例如“这是一件光学诱饵，我使用过。我们称之为电子水母。它只是 16 个蓝色 LED 灯，我们可以编程进行不同类型的[光]显示。”有趣的是，能够将当前研究中观察到的学习速率与 Table 1 中列出的研究中报告的学习速率进行比较，其中参与者在阅读后对词汇进行了更刻意的工作。无论如何，这样的比较都应该非常谨慎地解释，因为材料、程序、参与者等存在差异，但在本例中尤其困难，因为使用的测试格式多样。Table 1 中的一项研究，Hulstijn 和 Laufer(2001)，使用了与我们类似的意义回忆测试，并且也对部分正确响应给予认可。导致两次实验（使用不同的学生群体）中最佳学习增益的任务是撰写包含从注释阅读文本中 10 个目标词汇的文本。 阅读和写作任务总共花费了 70-80 分钟的时间。一次试验的延迟后测分数为 2.6，另一次为 3.7。我们在报告的实验中所记录的成功率与之相比并不逊色，考虑到实验仅通过大约 40 分钟的课堂程序完成。Lu（2013）也包括了一个意义回忆测试（但回答仅以二元方式评分）。学生们阅读了一个包含九个目标词汇的故事，然后执行了以这九个词汇为中心的任务，如包含这些九个词汇的总结写作（额外花费大约 40 分钟）和为每个单词完成三个空缺句子（额外花费大约 20 分钟）。平均而言，这两个条件下的学生在延迟后测中回忆了九个目标词汇中的两个。再次，我们在实验中观察到的学习，没有输入修改，也没有对学生进行刻意词汇学习的指导，似乎与之相比并不逊色。

As with other studies of vocabulary uptake from textual input, it needs to be borne in mind that the gains observed in a vocabulary test on a selected set of target words is likely to underestimatethe amount of acquisition that actually took place. After all, the participants may have acquired, fine-tuned, or strengthened knowledge of other words than the 18 target words we selected. Besides, there are of course mul- tiple other ways beyond word learning in which the participants are likely to have benefited from engaging with the TED Talk video and from the meaning-focused oral output activity.
与其他词汇吸收研究一样，考虑到从文本输入中获取的词汇量，需要记住，在选定的目标词汇集上进行的词汇测试中观察到的收益可能低估了实际发生的学习量。毕竟，参与者可能已经学习、优化或加强了我们选择的 18 个目标词汇之外的其他词汇的知识。此外，参与者从观看 TED 演讲视频以及参与意义导向的口语输出活动中很可能受益于多种方式，而不仅仅是词汇学习。

Even regarding the 18 target words we used to estimate vocabulary learning, it needs to be borne in mind that these were selected simply because they were unknown to the participants. Whether the studentsconsidered many of them to be of high utility and thus worthy of remembering is far from certain-but this is no different from numer- ous previous studies where target words were selected by virtue of their low frequency. It also needs to be acknowledged that we used only one test format to measure knowledge gains, and this was a rather challenging format. For example, if, apart from the meaning-recall for- mat, we had administered a less challenging meaning recognition test(in a multiple-choice format),this could have yielded evidence of learning that the recall test may have failed to reveal.
即使考虑到我们用来估算词汇学习的 18 个目标词汇，也需要注意的是，这些词汇之所以被选择，仅仅是因为它们对参与者来说是未知的。学生是否认为其中许多词汇具有高实用性，因此值得记住，这一点远非确定的——但这与许多先前的研究没有不同，在这些研究中，目标词汇是根据其低频率被选择的。还需要承认的是，我们仅使用了一种测试形式来衡量知识增长，而这是一种相当具有挑战性的形式。例如，如果我们除了意义回忆形式之外，还施加了一个更不具挑战性的意义识别测试（以选择题形式），这可能会提供回忆测试未能揭示的学习证据。

The current study focused on vocabulary uptake through an input- output-inputcycle, but using this cycle is likely to be beneficial in additional ways. For example, reviewing previous research in this area, Newton and Nguyen(2018)surmised that it can also help learners develop their L2 listening metacognition strategies. Specifically, when they are required to summarize listening content, learners need to syn- thesize information from their notes. This synthesis helps them notice gaps in their previous interpretation of input content, which, in turn, prompts them to direct their attention to relevant information as they listen to the same input a second time. Put differently, the input-out- put-input cycle provides learners with an opportunity to monitor their listening process, evaluate the listening outcome, and search for suit- able solutions to the listening problems they may have experienced(Vandergrift&Goh, 2012).
当前的研究专注于通过输入-输出-输入循环来获取词汇，但使用这个循环可能在其他方面也大有裨益。例如，回顾这一领域的先前研究，Newton 和 Nguyen（2018）推测它还可以帮助学习者发展他们的第二语言（L2）听力元认知策略。具体来说，当他们被要求总结听力内容时，学习者需要从笔记中综合信息。这种综合有助于他们注意到他们对输入内容之前的解释中的空白，这反过来又促使他们在第二次听取相同的输入时将注意力集中在相关的信息上。换句话说，输入-输出-输入循环为学习者提供了监控听力过程、评估听力结果以及寻找他们可能遇到的听力问题的合适解决方案的机会（Vandergrift&Goh, 2012）。

There are inevitably several limitations to the study we have reported here. One limitation is that only onetypeof audiovisual input and only one type of output task were put to the test. It is worth examining whether similar trends might be observed(and similar benefits reaped) in conceptual replications with other video materials(includingcaptioned ones) and with the implementation of other text-based output tasks, especially ones that stimulate learner-learner interaction instead of the unidirectional summary task tried here. A likely factor in the effectiveness of text-based output tasks for vocabu- lary uptake (when it is not made mandatory to use particular words) is the extent to which the learners will find words from the text useful or even essential to complete the output task they have been set. Even in the case of a summary or retell task, it is clear that non-key concepts(and associated words) may be avoided. In our data such words as bar- bel, mesh, and morsel were not recycled from the text and consequently stood a poor chance of being recalled later. Besides, as we have seen, even in the case of key concepts, learners may be able to resort to strategies such as substitution and paraphrase. Another limitation to this study is that it did not include another treatment condition where an output activity followed the initial viewing but without a subsequent viewing. This could have helped to determine to what extent the (announced) output task as such promotes vocabulary uptake, even in the absence of a second viewing.
我们在报告的研究中不可避免地存在一些局限性。一个局限性是仅测试了一种类型的音频视觉输入和一种类型的输出任务。值得探讨的是，在其他视频材料（包括字幕视频）以及实施其他基于文本的输出任务（特别是那些刺激学习者之间互动而不是这里尝试的单向总结任务）的情况下，是否能观察到类似的趋势（并获得类似的好处）。对于基于文本的输出任务在词汇吸收的有效性（当不需要使用特定词汇时），一个可能的因素是学习者将从文本中找到多少词汇有用，甚至必要，以完成分配的输出任务。即使在总结或复述任务的情况下，也很明显，非关键概念（及其相关词汇）可能会被避免。在我们的数据中，诸如胡须、网格和碎屑这样的词没有从文本中重复使用，因此后来回忆起来的机会很小。 此外，正如我们所见，即使在关键概念的情况下，学习者也可能能够采用替换和释义等策略。这项研究的另一个限制是，它没有包含另一个处理条件，在这个条件下，输出活动在初始观看后跟随，但没有后续观看。这本可以帮助确定输出任务本身在缺乏第二次观看的情况下，对词汇吸收的促进程度到底有多大。