英语学习中所谓“主动词汇”、“被动词汇”，很可能都是错觉（上）——（英语教学法原著选读70）
预告一下：因有事，本周五的教学法原著选读暂停一期，请朋友们见谅！
求助告示：请朋友们一定看到文章最后，并帮我点进广告，但不必关注、购买，进入广告页面后退出即可，不好退出的话可以连续点两次返回键，应该就能退出。每篇文章点一次即可。积少成多，系统会给我一点广告费。感谢你们的大力支持！
------------------------
“故老相传”，我们学英语的时候，学到的词汇分为“主动词汇”和“被动词汇”两类，其中被动词汇在输出过程中往往用不起来，而主动词汇则是那些能够在输出性活动中能够熟练使用的词，感觉上是我们在主动使用它们，故名。在词汇教学研究界也有这样的说法，并且还有人煞有介事地进行了这方面的“估算”和“测算”，有说被动词汇量是主动词汇量两倍左右的，也有说是五倍多的，还有说被动词汇量能达到二十多万的，众说纷纭，也不知道谁说的是真的。
不过，不管真还是不真，国内一些英语老师那是拿来就用，一本正经地跟学生说，被动词汇量是主动词汇量的两倍，要记住哦！我就一直受到这样的教育，还曾经沾沾自喜：我在网上测词汇量，得知自己有24000词的被动词汇量，想想已经离莎士比亚40000词的数量不远了，至少，过半了嘛！
好在学术界还是有人会较真的。在上海外语教育出版社引进出版的《教学法丛书》中有一本《词汇：描述、习得与教学》（Vocabulary: Description, Acquisition and Pedagogy），收录了关于词汇研究的近20篇论文：
这本书里的第五篇，“Receptive vs. productive aspects of vocabulary 词汇的输入与产出两方面”，作者Francine Melka，探讨的就是究竟有没有“主动词汇”、“被动词汇”、如果有、究竟又怎样的特征等话题。下文中的大写R指代Receptive vocabulary，大写P指代Productive vocabulary，请朋友们注意。另外，在下列文字中，以下几对术语是基本同义、可互换使用的：
receptive vocabulary -- productive vocabulary
输入性词汇 -- 产出性词汇
passive vocabulary -- active vocabulary
被动词汇 -- 主动词汇
production -- reception, comprehension
产出 -- 输入，理解
作者Melka的论文主要是沿着这样的逻辑路线进行探讨的：有各种“主动词汇”、“被动词汇”的数字估算，那么这些数字是怎么测算出来的呢？变量控制做得怎么样？在对前面的多项“主动”、“被动”词汇量测算进行了考察之后，作者对各家不同的“主动”、“被动”词汇量之比（如有的说一比二，有的说一比五，如此等等）进行了思考和验证，下文就是作者对这些各异的测量结果之间的偏差做出的解释。本节标题“Reasons for discrepancies between estimates of reception and production”，输入词汇量与产出词汇量各种估算数据间出现不匹配的原因，主要说了六个方面：
1、测试所用的词汇类型。如果测试的词汇样本中常用词居多，那么这些词必然同时属于主动词汇和被动词汇两类，则测算出来的主动词汇量和被动词汇量之间的差距就不会很大，连两倍关系都达不到；相反，如果生僻词居多，则必定是主动词汇量很小、被动词汇量很大——勉强认识很多但基本不会用嘛。
2、测算实验中的试卷批阅尺度。一般而言，测试主持者对被动词汇量的认定比较宽松，对主动词汇量的认定则比较严格，因而会出现被动词汇量比主动词汇量大得多的情况。如果反过来，则两者之间的差距就不会很大了。
3、上下文所起到的作用。一般来说，较为充分的上下文不管是对主动词汇的使用还是对被动词汇的识别，都有很显著的促进作用。举一个简单的例子：有时我们从卧室里急匆匆地走过客厅、走进厨房，想做点什么，但等到进了厨房却忘记了。这时很多人会选择倒回到卧室里去，因为那里的环境会帮助我们回想起最初我们是想做什么的。从语言的角度说，这就是上下文——或曰语境——对产出所起到的促进作用。
4、有意无意地避开某些词汇。有时受试者是因为文化原因避免使用特定词汇，比如禁忌语，也有时是因为发音、拼写没把握而不去使用。不管怎样，避开了这些词汇，就使得这些词汇从主动词汇中消失了。这种情况在被动词汇的测算中基本上是见不到的。
5、可能使用的词与实际使用的词。人的性格不同，在词汇的使用上也有不同。有的人胆大，学过什么词就试着使用；有的人则小心谨慎，没把握的词不用，特别是在有外部压力——比如有大批听众——的时候，更是如此。这对主动词汇的测算也有不小的影响。
6、母语和同源词（指的是两种语言里的对应词，比如日语的“家族”和中文的“亲属”）。在二语学习的初期，如果母语和二语同属一个语族，比如荷兰人学习法语，则会有相当数量的同源词，母语这时起到的是正迁移作用。然而，到了后期，学习者会了解到所谓的“同源词”只是在词义的一些方面相同或近似，需要掌握的不同情况还有很多，这时他们往往会避免过多使用同源词。这对主动词汇量的测算也是一个影响因素。
总之，主动词汇和被动词汇之间并没有我们想象的那样有巨大的数据差异，研究也日益表明，并不存在两个相互独立的、各自运行的词汇系统。
------------------------
下周一我将发本篇的（下），探讨的是词汇学习的连贯性，并从本质上对所谓的“主动词汇”、“被动词汇”作出解读。敬请期待！
我经常在个人微信（1601918196，这也是我的QQ）上探讨英语学习与教学，欢迎朋友们添加，共同探讨。
A certain number of factors linguistic and extra-linguistic seem to affect estimates of R and P and to give either more importance to the receptive vocabulary or to the productive one depending on the way a test is set up or the results are interpreted. the following factors are not exhaustive, but indicate a number of critical points (see also Melka Teichroew, 1982).
i Type of words
Many vocabulary tests are based on dictionary samples or vocabulary lists. The problem of choosing good dictionary samples is related to the size of the dictionary and also to the range of difficulty of words: the choice of tested items explains the (over)estimates of vocabulary size of studies like Holden (1890) and Seashore and Eckerson (op. cit). It seems that ‘simpler' words, i.e. basic words highest on a frequency list, are typically both recognized and produced by subjects. This explains the relatively small gap between receptive and productive vocabularies in Seashore and Eckerson's tests; their lists were composed of 'common basic words' (which were both recognized and produced), plus a very small percentage of 'rare words', which the subjects neither recognized nor could produce. The choice of these words (many basic words and few low-frequency words) probably affected the vocabulary estimates: the gap between R and P was small since the words were either understood and produced or neither understood nor produced.
Nation (1993) proposes various criteria to choose good representative samples from a dictionary. The author reviews dangers and problems encountered by makers of tests in using dictionary samples: these dangers have to do with the choice of the sample source, the decision to include or not include derivatives in the list, the selection of high frequency words, and the way of checking representativeness of a sample list. The use or misuse of these criteria could explain, among other things, the overestimation or underestimation of vocabulary sizes.
2 The grading of a test
Another factor which could influence the estimations of R or P vocabularies has to do with the way tests are graded and the criteria to decide whether an item is considered correct. In the receptive tests used in many of the above studies, a word was considered 'known' when a subject was able to produce the slightest correct element of meaning. Some researchers are very undemanding with regard to criteria: Hartmann (op. cit.: 437) accepts, in receptive tests, almost anything provided the answer indicated the testee had the slightest familiarity with the real meaning of the word. The same is true for productive tests; Seashore and Eckerson (op. cit.) ask subjects to give all meanings they know for a word and grade accordingly. If R tests are marked leniently and P tests marked strictly, the resulting gap measurement is large. If the opposite is true in marking, then the gap will appear smaller (or even in favour of production).
3 The role of context
Related to the procedure in setting up or designing a test, and indirectly to the way of grading it, context plays an important role in explaining the gap between R and P. In multiple-choice and translation tests, the linguistic context is rather important: it can be very restrictive or can 'give away' a word. In the case of productive tests of spontaneous discourse or speech, context, linguistic or extra-linguistic, is also very important because it helps the subject retrieve a word, though Morton (1979: 115) suggests that 'context operates in the same way whether we are generating or recognizing a stimulus' (see also Read，1993: 357).
It is also clear that context greatly aids comprehension. Linguistic context, which includes the ability a subject has of deducing morphological rules or lexical occurrences rules, facilitates understanding of a word. Likewise, extra-linguistic context, induding notions of culture, use of situations, gestures, etc., also helps better understanding, although it can also give the deceptive impression of a greater receptive knowledge than actually exists, as mentioned above (Clark et al., 1974, Clark, 1993).
Nagy (1.4) suggests that, apart from explicit instruction, people pick up much of their vocabulary knowledge from context. The strategic use of context can bring the subject to the highest level of comprehension; Nagy reports that an average student could 'learn' to recognize up to 1,000 words per year from reading material. He also suggests that, at first, context plays a less important role for L2 learners because they have not yet achieved a high level of L2 proficiency. Later though, context plays a bigger role. The use of context is, then, a crucial strategy for dealing with unfamiliar words L2 learners encounter when reading texts.
In an experiment with university students, Van Koppen (1987) compared R and P vocabularies. The main variable was the context. Her results show the importance of context (versus no context) and that context has a similar and facilitating effect on receptive and productive results.
This section shows that context plays a role in estimating R and also P, even though some authors seem to believe intuitively that context primarily helps receptive knowledge. Context could, in some cases, have the effect of indicating an overly large gap between R and P.
4 The notion of avoidance
Another explanation of the gap between R and P in tests may be found in a pragmatic phenomena more or less controlled by subjects--the notion of avoidance. A subject can consciously refuse to use known words (taboo words, for example) for extra-linguistic (primarily cultural) reasons. This is what Blum and Levenston (1978) call ‘true avoidance’. 'Apparent avoidance’ is superficial; it is a refusal or reticence to use a word not well known because of speliing or pronunciation difficulties, or became of uncertainty about its grammatical or semantic features. Avoidance, then, becomes a negative factor for production.
5 Possible words versus actual words
Some people seem to use all the words they know, some do not. This raises the problem of 'possible use' versus 'actual use'. Actual use includes common words one knows well and uses regularly (see Aitchison's image of well worn paths for well-known words, op. cit.: 195). 'Possible use' would be a reserve of words which may be just barely productive, or which do not always surface productively, but do surface on some occasions depending on individual variation, culture, education, as well as on social or psychological circumstances, i.e. feeling secure in front of an audience (see Terrell et al., 1977). For these 'possible words', Aitchison uses the metaphor of ‘narrow and dimly lit paths’（op. cit.: 195}，when words are forgotten, but could eventually be used. Depending on how one considers ‘possible use’, whether a part of receptive knowledge or productive knowledge, one category or other will be inflated and the gap between R and P will thus be affected.
6 The role of L1 and of cognates
The learning of some L1 items seems to require less effort of foreign language learners when their native language and L2 belong to the same language family (Holmes and Ramos, 1993: 87).
The cross-linguistic equivalence is all the more easy to perceive when L2 items formally resemble L1 items (more or less cognates), in spelling, morphology and/or syntax: foyer in French and Dutch, reception/reception in French and English. Learners, then, anticipate that the meaning of these items is approximately similar, which is the case in the above examples. The equivalence assumed by the learner in the case of the polysemous foyer in French, for example, in the first stage of learning, helps the learner in the phase of recognition of words by oversimplifying; the learner assumes that the L1 item and the L2 item are equivalent. It seems that, at this stage, cognates will be important in helping recognition (see Meara, 1993: 283; Palmberg, 1985).
In the course of time, learners realize that the equivalences are more limited than previously thought and they modify the 'inadeqiquacies of these equivalences perceived initially' (Ringbom, 1983:165). Foyer means a lot more in French than in Dutch; a Dutch learner of French has to assimilate all the other polysemous meanings in order to consider foyer as a known word. In production, it is possible that a learner could hesitate to u$e a cognate for which a phonetic or graphic resemblance with an L1 item does exist but which he or she knows has (at least partially) different meaning(s) in the L1 and L2 (see the paragraph concerning avoidance).
In principle, establishing equivalences between items in an L1 and L2, i.e. discovering cognates between the two languages, is part of language competence, and thus should serve production as well as comprehension. However, the learning stage of the L2 learner is important. As a beginner, the L2 learner has a tendency to generalize equivalences and to use this principle in comprehension and production. In later stages of learning, the same learner hesitates to produce cognates whose meaning he or she is not sure about. Reception will be maximal then, and production will be avoided. Thus, cognates can serve to give the appearance of increased or decreased productive capability, depending on where one is at on the learning plane. This again affects the perceived size of the gap between R and P.