Exploiting Lexical Properties for L2 Students of Chinese
Jerome L. Packard
Abstract This paper proposes lexical coding strategies that follow from experimental results demonstrating that the component of working memory known as Focal Attention （FA） is as robust in learners of Chinese as it is in native Chinese speakers. Lexical encoding that relies upon FA memory can be strategically employed by L2 learners of Chinese to improve the efficiency of their lexical storage and retrieval.
Keywords Chinese L2 vocabulary；focal attention；lexical storage and retrieval；working memory
1. Introduction-FA as a Critical Memory Component
This paper addresses the question of how we can improve Chinese second language learners’ lexical storage and retrieval capability. This is an especially germane question given the intuition that second language students may appear to have a limited verbal working memory capacity in their L2 compared to their native language. The goal of this paper is to discuss how to utilize L2 learner memory capacity in a way that will make lexical information easier to both store and retrieve.
The basic strategy is to exploit the specific characteristics of L2 working mem-ory as a means of making L2 vocabulary easier to process. One way to do this is when L2 learners encode or memorize lexical items, to have them utilize information that is to be incorporated within the domain of the Focal Attention （FA） component of working memory. In this way the locally-attached information remains within the scope of FA working memory because the information is directly attached to the lexical item where it can be directly processed.
Scholars who study memory are of different minds regarding what constitutes the content of FA, but they do generally agree that FA is that part of memory that requires the least cognitive effort. The results of the experiment of Packard and Qian （2016） support the view that the part of working memory that is not as good for second language learners is ‘non-FA working memory’ while the part of verbal working memory that is virtually the same as for native speakers is the FA component, consisting of the first 1500 milliseconds or so of speech perception. One of the results of that experiment is that second language learners’ ‘non-FA working memory’ was found to be weaker than that of native speakers, but that second language learners’ ‘FA working memory’ is equivalent to the FA working memory of native speakers of a given language. Therefore it is the non-FA working memory that is limited in L2 speakers, and it is the ‘recent’, FA component of working memory that is not limited in L2 speakers. In other words, the FA capacity of second language learners of a given language appears to be virtually the same as for native speakers of that language （Packard and Qian 2016）.
For this reason, because FA in L2 learners is as robust as in L1 speakers, we should use the FA memory component as much as possible when engaged in lexical encoding. The model used here to illustrate memory operations is adapted from the version described by McElree （2006） as seen in Figure 1. The memory component is depicted by three concentric circles, with the largest circle on the outside represent-ing long-term memory and the two circles on the inside representing working memory and Focal Attention respectively. The circle representing FA within working memory is the part of memory that is active and accessible within the conscious awareness of the L2 speaker. It is this FA memory component that may be focused upon and utilized by L2 language learners.
The result of the experiment of Packard and Qian （2016） was that the second language learners’ FA memory was just as strong as native speakers’ FA memory, while second language learners’ non-FA working memory was less strong than the non-FA working memory of native speakers. These investigators found that second language working memory has specific limitations but that the FA component is just as strong in second language speakers as it is in native speakers. Therefore if we assume that FA memory is just as strong in second language learners as in native speakers, an optimum strategy for L2 learners may be to try to utilize the power of FA working memory to make L2 vocabulary easier to store and retrieve.
One way to make L2 vocabulary easier to process is to have L2 learners equip vocabulary items with locally-attached information that is limited in size and processed within the domain of FA when the vocabulary is encoded. In this way, the locally-attached component of vocabulary information is within the scope of FA because the information is attached as part of the lexical item and as a consequence does not take up large amounts of storage space or processing time but yet is more efficiently processed. As discussed in McElree （2006）, it is possible to represent memory information in three possible states, either in LTM, WM, or in the current FA （see Figure 1）, with each state associated with a distinct retrieval speed. Information that resides in LTM would be associated with the slowest retrieval speed, information within the span of WM would be faster than LTM but slower than FA and information within FA would be retrieved fastest of all.
Figure 1 A Model of Memory containing a Focal Attention Component
（modified from McElree, 2006）.
The hypothesis being advanced here is that if information consists of one item within FA when encoded, it will be accessed more quickly when retrieved. The relevance for learning vocabulary is that the entry of information into the lexicon should be attached directly to the lexical item being encoded, because if it is at-tached to the lexical item at the time of encoding then it is attached while the lexical item is within the span of FA, and therefore initially stored during the period of time that FA is active, making it possible to retrieve it more quickly because it was within the span of FA processing when encoded.
The critical point here is that lexical encoding may represent the type of complex cognitive processing that depends on our ability to rapidly shunt information between FA and the other components of memory. According to this model （McElree 2006, p.177）, vocabulary acquisition （lexical encoding） and retrieval may not rely upon a temporary memory store such as WM, but rather may depend more on our ability to shunt information rapidly between FA and LTM, with vocabulary retrieval requiring the restoring of passive representations （i.e., the lexicon） outside of FA to active processing.
2. Two-item lexical FA encoding
Typically the information processed in the hearer’s FA will be the last item processed by the hearer. However, the contents of FA may consist of more than a single item if multiple items are encoded into a single ‘chunk’, in which case a single ‘chunk’ that contains more than one item may occupy FA. This is relevant to the lexical encoding of Chinese, because it might entail a two-morpheme Chinese lexical item occupying FA, causing the two-morpheme lexical item to be easier to encode and retrieve because it is located within FA. Under these conditions, not only do the vocabulary items being memorized possess the identities of ‘word’ — that is, words are the items that are to be listed into the lexical store — but they are also two-item chunks whose acquisition occurs within the span of FA, therefore making them easier to encode and retrieve. The specific hypothesis being advanced here is that an item encoded in FA will be retrieved more quickly because it was in the time-span of FA that it was initially encoded.
To give an example, imagine an item [A-B] （e.g., chudian contact-electricity触电 ‘to get an electric shock’） encoded as a unit within FA, compared with items [A and B] （e.g., chu触 ‘contact’ and dian电‘electricity’） encoded as single, separate units within FA. The question would be whether, e.g., the morpheme dian电‘electricity’ would be more easily retrieved under the [A-B] or the [A and B] encoding condition. The learner’s ability to store and retrieve dian might be more efficient in the [A-B] condition than in the [A and B] condition, because under the former condition, both dian and its lexical contextual partner chu were encoded within the span of FA, while under the latter condition, only dian — and not chu — was encoded within the span of FA.
The advantage of FA encoding of chudian for the retrieval of dian lies in the fact that the two-item （two-morpheme） chunk （word） is encoded within the span of FA, with dian containing the added, contextual information of the presence of ‘chu’. Furthermore, the morpheme dian in additional two-morpheme lexical forms such as dianying electricity-shadow 电影 ‘movie’, tingdian stop-electricity 停电 ‘power failure’ and dianshi electricity-view 电视 ‘television’ is better remembered and more easily retrieved because dian is encoded multiple times, with each instance of encoding being contextualized differently, but still occurring within the span of FA. This takes advantage of the fact that children learn to read and write Chinese characters better when their number and variety of morphological contexts is increased （Packard et al. 2006）, and also utilizes the fact that the target morphemes within the word fall within the scope of the FA component.
In this way, the lexical collocation information of the morpheme dian is util-ized by introducing it within several different words that contain the morpheme in different positions within the word. The result is that when this target morpheme is taught it can include different collateral lexical information occurring both to the left and to the right of the target morpheme. Learners of Chinese vocabulary are thus able to take advantage of the power of collocation information that naturally occurs in both the morphology and the syntax of Chinese lexical items （see, e.g., Packard 2018）. According to general tenets of learning theory, imbuing data to be stored and retrieved with greater varieties of accompanying contextual information should increase the representational strength of the data entry, thereby making it easier to process and retrieve.
To summarize, the previous examples using dian illustrate the power of multiple collocations of word components that provide multiple contexts for the encoding of word and morpheme information. In order to facilitate the storage and retrieval of lexical items, we can use multiple sources of morphological information to serve as multiple storage and retrieval encodings within FA that will allow those lexical items to be more efficiently and reliably processed.
3. Effects of directionality in incremental build-up repetition drills
Another way to effectively use the FA portion of memory is to ensure that the target vocabulary item remains within the span of FA during speech processing and production. One way to do this is to have the target vocabulary item remain in that portion of the repeated phrase that is located within FA. For this reason, in so-called ‘build-up’ drills — in which items are incrementally added word-by-word to the phrase being built up — if the ‘build-up’ drill operates in the ‘backward’ direction, it is possible to keep the target item within the span of FA. Since in the case of ‘backward build-up’ the repeated item is the ‘last to be produced’ by the speaker and the ‘last to be heard’ by the hearer, the target item constitutes the content that occurs within FA for the hearer over multiple iterations. In ‘backward build-up’, the contents of FA are continuously and iteratively updated, increasing the likelihood that those specific FA memory contents will be retained.
To give an example, let us assume that the target item to be drilled is the Chinese word wanfan late-food晚饭 ‘dinner’. In an incremental build-up drill, to keep wanfan in the position of FA it needs to be at the end of the phrase and the incremental build-up has to be in the ‘backward’ direction, that is, starting with the last word of the phrase and incrementally adding elements from ‘right-to-left’, moving from the end of the phrase to the beginning, until the entire phrase is produced. This would occur as follows:
backward build-up
‘I don’t want to eat dinner.’
forward build-up
‘Dinner, I don’t want to eat.’
It is clear that in the ‘backward’ build-up condition, the target word wanfan is repeated with each iteration, and in addition with each iteration the target word is not only repeated, it is also repeated in the position of ‘recency’, i.e., it is the last item produced, which places it within the span of FA with each iteration. In the ‘build-up’ drill, ideally, the goal is to keep the target item in the position of ‘recency’, which would be the position of FA.
In contrast, in the ‘forward’ build-up condition, since the target word is top-icalized, it is repeated with each iteration, just as it is in the ‘backward’ build-up condition. A major difference, however, is that in the ‘backward’ build-up condition the target word wanfan occurs within the span of FA with each iteration because it is the last item produced in each iteration, while the target word in the ‘forward’ build-up condition is the first word to be produced in each iteration, allowing it to fall outside the span of FA memory. In addition, since ‘backward’ build-up has been shown experimentally to result in fewer errors produced by students than ‘forward’ build-up （Li 2018）, the use of ‘backward’ build-up results in both optimum memory utilization and optimum production performance.
4. Conclusion – Utilizing Lexical Properties and Focal Attention
The take-away message of this paper is that teachers and learners of Chinese as a second language have at their disposal a working memory component known as Focus of Attention that is as robust in second-language learners as it is in native speakers, and therefore should be utilized to the greatest extent possible. In order to most effectively accomplish this, two tactics can be pursued in the L2 teaching and acquisition of Mandarin. The first is to select two-morpheme vocabulary items as targets for acquisition, and the second is to manipulate those items in a manner that keeps them within the span of FA. The first goal can be accomplished by selecting for acquisition a variety of two-morpheme words each of which contains the given target morpheme, and the second goal can be accomplished by implementing activ-ities such as ‘backward build-up’ that keep the two-morpheme target words located within the span of FA memory.
Li, You（2018）Directionality Effects on L2 Sentence Build-up Drills in Mandarin. Paper presented at the 30th North American Conference on Chinese Linguistics, Columbus, OH.
McElree, Brian（2006）Accessing Recent Events. In B. H. Ross （Ed.）,The Psychology of Learning and Motivation, Vol. 46: Advances in Research and Theory （pp.155-200）. San Diego, CA, US: Elsevier Academic Press.
Packard, J.L.（2018）Memonics in the Chinese L2 Lexicon.Taiwan Journal of Chinese as a Second Language, 15（2）,53-64.
Packard, J. L., X. Chen, W. Li, X. Wu, J. S. Gaffney, H. Li, and R. C. Anderson.（2006）Explicit Instruction in Orthographic Structure and Word Morphology Helps Chinese Children Learn to Write Characters. Reading and Writing 19（5）:457-487.
Packard, J. L. and Z. Qian.（2016）A Working Memory Explanation for Recency Effects in Mandarin Second-language Sentence Processing.《世界汉语教学》 （Chinese Teaching in the World）, 30（1）:75-100.
（Jerome Packard, East Asian Languages and Cultures,
University of Illinois at Urbana-Champaign / U.S.A., jpackard@illinois.edu）