Yesterday evening I experienced a wonderful lesson in Lakota language taught by a grad student in the Linguistics department at the University of Hawai'i (I will add her name here if she wants me to). This was her second attempt at CI teaching, and she excelled at going slow, parsing, speaking clearly, presenting a clear word list with pictures, pausing and pointing, asking questions in the target language, integrating culture in popups, focusing on form in popups, engaging the audience with questions, fishing for audience contributions, and more fine-grained practices that I'm not noticing here. At one point, an audience member asked, "Can I get a popup? You keep saying 'The dog, the cat, likes. Does that mean the dog likes the cat or the cat likes the dog?'" Clearly, pointing to the pictures of the dog and the cat in the order spoken was not clear enough to many people in the room, to clarify "who likes whom". This was important information for our unfolding, co-created scene, and audience participation could not continue if people weren't clear on what we had all decided to be true. I jumped in and mentioned a relevant finding in Bill VanPatten's research on Input Processing (IP; VanPatten & Cadierno, 1993), showing that learners tend to assume that the first noun is the agent or subject of the sentence. But the audience still wanted a clear mention of the intended sentence meaning. Upon reflection after this lesson, I realized that the key word in research is "tend to."
Experimental researchers are rarely concerned with unanimous outcomes, that is, every member in one group coming out equal. Instead, researchers seek out tendencies in averages. So if a teacher is aiming for equity among learners, then it's important to understand that averages found in research only imply averages in our classrooms. Exceptions to the average should also be implied from the research. So, if a teacher's goal is success for everyone, then findings from research should only serve as a helpful starting point (the teacher can think, "Ok, many students will likely assume the first noun is the agent or subject in the sentence"), but then the teacher should differentiate by checking for which participants are similar to the non-average members in the research. In this example, a teacher can still help many people in the room to more clearly understand the unfamiliar language structure by asking, "What do you think this sentence means?". The teacher can then briefly clarify using a shared language, and then return to the discussion in the target language (this is standard practice in the TPRS literature). In addition to facilitating language acquisition (connecting form with meaning frequently), these extra clarifications should help audience members (the learners) contribute to the ongoing discussion as they are more clear on what is being discussed (the content of the discussion).
0 Comments
I made a new card-grid page for CI teachers to hold and look at, for reference when asking actors and audience members questions during 'story-asking'. I imagine it could be used or adapted for other CI activities as well (PQA, Special Person, etc.). If teachers want students to ask more questions, they can also project this for learners to see. More is explained on the first page in the document for download, below.
It's normal for teachers new to CI (and experienced ones as well) to hear an internal voice saying: "That's too much repetition! The students are bored for sure! Move on!" In response, I want to illustrate a difference between repeating language forms and repeating situated meaning.
FIRST, JUST START READING: FROM WIKIPIEDIA (https://en.wikipedia.org/wiki/Shaggy_dog_story) Archetypal story A shaggy dog, the archetypical subject of long-winded, pointless stories The commonly believed archetype of the shaggy dog story is a story that concerns a shaggy dog. The story builds up, repeatedly emphasizing how shaggy the dog is. At the climax of the story, someone in the story reacts with, "That dog's not so shaggy." The expectations of the audience that have been built up by the presentation of the story, that the story will end with a punchline, are thus disappointed. Ted Cohen gives the following example of this story:[1] AT THIS POINT IN THE READING, THE WORD "OF" HAS APPEARED 8 TIMES (INCLUDING IN THIS SENTENCE), AND THE WORD "THE" HAS APPEARED 21 TIMES. THAT'S PROBABLY ENOUGH, AND I'M AFRAID YOU WILL FEEL BORED, SO I REMOVED THESE TWO WORDS FROM NOW ON. A boy owned a dog that was uncommonly shaggy. Many people remarked upon its considerable shagginess. When boy learned that there are contests for shaggy dogs, he entered his dog. dog won first prize for shagginess in both local and regional competitions. boy entered dog in ever-larger contests, until finally he entered it in world championship for shaggy dogs. When judges had inspected all competing dogs, they remarked about boy's dog: "He's not that shaggy." THE WORD "SHAGGY" HAS APPEARED 11 TIMES, AND "STORY" 10 TIMES. PROBABLY ENOUGH. ALSO "IS" (8 TIMES ALREADY). I'LL STOP USING THEM SO THEY DON'T BOTHER YOU. However, authorities disagree as to whether this particular archetype after which category named. Eric Partridge, for example, provides a very different story, as do William and Mary Morrin Morris Dictionary Word and Phrase Origins. NOW LET'S STOP WITH "A", "AN", AND "TO". YOU'VE HAD ENOUGH OF THEM, I THINK. According Partridge and Morrises, archetypical dog involves advertisement placed in Times announcing search for dog. In Partridge story, aristocratic family living in Park Lane searching for lost dog, and American answers advertisement with dog that he has found and personally brought across Atlantic, only tbe received by butler at end who takes one look at dog and shuts door in his face, saying, "But not so as that, sir!" In Morris, advertiser organizing competition find dog in world, and after lengthy exposition search for such dog, winner presented aristocratic instigator competition, who says, "I don't think he's so."[3][4] My point here is that in order to combat that irrational voice in our heads, we should recognize that language forms can be repeated infinite times if its users (listeners, readers, speakers, writers) are focused on the meaning in each situation. This situated meaning keeps changing and progressing logically as we work through our stories, personal info, opinions, experiences, etc., and so everyone can remain engaged in the discussion. When reading, we just wanted to know if Oprah will get the Sriracha, if the worms in my backpack were bloody or not, and if Dan's house had rats or not. Other people would want to talk about other things, and that's the challenge for the teacher: keep offering learners topics to talk about until we show interest. Then talk about that. One discussion that comes up from time to time among CI teachers is how to choose the language chunks, or "items", or "structures", that we teach. Martina Bex has a very useful description of "language chunks" on her blog here. In addition, for extra perspective, I would like to describe here what Usage-based Linguistics researchers call "constructions".
Constructions are not simply language written on a page, but are a psychological construct, measured in the human mind through corpus research. A construction is defined as any linguistic symbol of any length that the mind connects with meaning. A word is a construction because it is a sound (or written string of writing, or sign language signs) that represents meaning. A morpheme is the smallest linguistic unit that represents meaning, and so it is also a construction. The more interesting constructions in the Usage-based Linguistics literature are those that help explain grammar in the mind. As VanPatten often says on his podcast, Tea with BVP, language is shaped by Universal Grammar and general learning architecture. Usage-based linguists seeks to better understand this general learning architecture, holding to what we call the "cognitive commitment". This means assuming that Universal Grammar is not real, and instead looking to exhaust all possible other learning mechanisms before allowing for which parts of language may indeed be shaped by Universal Grammar. Usage-based linguists seek to explain grammar as general pattern recognition and human categorization processes, which will be explained below as we look at phrase-length constructions. To be clear, both Usage-based AND Universal Grammar models of language view input as central to language development. Neither of these research camps argue that learners should learn an explicit rule, and then practice that rule, nor should they use problem solving abilities to master a rule. Input as raw language data (hearing and reading the language itself) is still necessary and central to acquisition in Usage-based models. So what is a multi-word construction? A multi-word construction consists of multiple slots in a particular order, that together represent meaning. Each slot can be either fixed, or open but limited to certain categories. For example, "wants to X" is a construction with two fixed words and an open slot. The open slot in this construction is limited to the verbs category. Another example is "put X Loc", where "put" is a fixed word, the X slot is limited to grammatical objects (the cat, the car, a fork), and Loc = locative phrase, is composed of a preposition (on, into, under) and a location (the table, the cave, the bed). So a construction contains not only the known words, but also the open slots for a limited selection of other words. Corpus research has shown that some words are more common in each construction than are other words, but that's for another post. Does it make much difference whether teachers plan to input only fixed "language chunks" or we plan for the fixed words plus open slots together in constructions? Maybe not, because teachers will fill in those open slots anyways when giving input to students. But human categorization needs exposure to a limited number of different examples in those slots before a category unconsciously develops in the mind. So if we are going to teach "sees a movie", the acquirer's mind is more likely to unconsciously build a category of "things we can see" in the construction's X slot (in a "sees X" construction), compared to if they only hear "sees a movie" (all words fixed) hundreds of times. This is one of the beautiful features of Circling, in which the slot openness is automatically built into the input. There is a lot more to be said about Cognitive Science, Cognitive Linguistics, Usage-based Linguistics, and Construction Grammar. I recognize that I have done them all very little justice in writing such a short post, but I hope this basic concept might offer a new perspective on the language we choose to offer our learners in their input. At first I wanted to title this post "Differentiating for Heritage & Non-heritage Learners", but my idea is really the opposite, as I will explain.
In my CI/TPRS teaching, I design my instruction to promote language proficiency by providing input in the context of communication ("The interpretation, expression, and (sometimes) negotiation of meaning in a given context", to borrow Bill VanPatten's most recent definition of communication). One central way I do this is by ensuring that the language learners hear and read contain words and structures (the input) which are repeated a lot (a "high frequency in the input" to use the SLA/Usage-based research term I like). The standard TPRS practice is to get this repetition by Circling, where the teacher asks lots of similar questions slowly, so our beginning learners have opportunities to process the same words in a variety of target-like sentence contexts. When I hear Vietnamese and more advanced French, Spanish, and German in teaching demos, I am usually eager to hear even the same sentence a few more times. It's at my current ability to process and my current level of challenge. When teaching Mandarin to undergrads at the University of Hawai‘i, and to 2nd, 3rd, & 4th graders at Punahou School's after-school immersion program, I'm always teaching heritage learners (often half to 3/4 of my class) who already understood me the first time I ask a question. Circling questions are confusing to them. They understood me the first time, so move on. But the non-heritage learners want to hear it again. My solution for balancing everyone out, and keeping us all in the class discussion during the listening part of the lesson is what I'm calling "Bottomless PictureTalk" (basic PictureTalk is described here, and please message me if this idea has already been described elsewhere). If I can ask two questions about each picture and three about how the students' own lives relate to each picture, and I do this for 20 separate pictures, then that's already (5 x 20) 100 times learners hear and process a target construction (or an otherwise emergent construction generated by the students when I'm doing non-targeted CI teaching). "But wouldn't many students feel annoyed by 100 similar questions?" To answer to this question we can simply do a web page search. I recommend trying this now: 1) open a long web page, maybe the Wikipedia page for "China", 2) search the page for the word "the", 3) search for "of". My search found over 1000 matches for "the" and another 1000 matches for "of". But shouldn't I have felt annoyed at so many uses of "the" and "of"? Probably not, because language naturally repeats when we keep talking about different things. If you go shopping with someone, you might ask each other "What about this one?" as many times as you need until you find what you needed. Neither person is likely to say, "Okay already!! Please use a different phrase. I'm so tired of this one." We use language to focus on meaning, and if the referent keeps changing, we don't notice the repeated use of language. My thinking now is that when we are all talking about new things, new people, new places, and new events, we don't have to differentiate for heritage and non-heritage learners. Everyone can continue being engaged in, and contributing to, our discussion in the target language. The non-heritage learners get the repeating language as we talk briefly about each picture, and no one suffers from repeated content. See how I make my long PictureTalk files as speedily as possible at this post here. |
Reed Riggs (Author)
Archives
April 2022
Categories
All
|