Different ways to model input

In this post I'd like to explore a range of ways in which input language can be presented and modelled with students. I'm starting from the assumption that when we want students to learn and eventually internalise (some would say 'acquire') new vocabulary and grammatical patterns, it makes sense to give meaningful samples of language at the start of a learning sequence. So, I shall offer some examples of how we can model input and comment on possible merits (and maybe disadvantages) of each. I'm also working on the assumption that the input language needs to be largely or wholly comprehensible. This is a slightly problematic issue, since input can be made instantly comprehensible by providing glossaries, translations and even whole parallel text translations. This means that input may contain many new words and chunks. In addition, students vary, so what is comprehensible to some, won't be for others. 

Written texts

For near-beginners and intermediates the written text in short or paragraphs is a pretty standard way to present language. New vocabulary can be made comprehensible by providing glosses (usually translations) at the end of the text, in the margin or even in brackets within the text. In theory the closer you can place the translation to the L2 text, the easier it is for students to register it and the more likely it is students will actually use the glosses. Pictures are sometimes added to support meaning.

Although texts of this type may not contain lots of examples of the same vocab and patterns, recycling of these features can be incorporated into the teaching sequence through question-answer ('circling') and lots of other drill-like interactions.

This is a pretty classic way to introduce and model language and does allow for a certain degree of authenticity, although, as with the other examples here, the type of input students get is what researchers call modified or simplified. There are excellent reasons to simplify input so that it aligns most closely with the current level of students knowledge and ability to pick up new patterns.

Note that with narrow reading texts, you can build in many more repetitions of patterns and vocab, since each paragraph can recycle the same items. This is called input flooding, the theory being that with more repetitions of items, the more likely it is they will be recalled and reused.

By the way, traditional written texts can include valuable cultural input.

Sentence builders

These are that type of substitution table used typically in the Conti-style EPI method (Extensive Processing Instruction). Gianfranco himself points out that this is his favoured tool for modelling input, but that others can be valid The grids contain words and chunks in columns and rows, with each item usually being translated into L1 (e.g. English). Not all L2 items have to be translated. If you are certain that the class knows some of the input already, you might think it's a waste of time to translate everything.

This clarity and total comprehensibility could feasibly be seen as a disadvantage, if you believe that students should be presented with the opportunity to infer or work out some meanings by other means. There is some research support for making students work a bit harder (deeper processing = better memory) - creating so-called desirable difficulty. Thus, you might make the case for adapting your sentence builder to the needs of the class. weaker classes might benefit more from that 100% comprehensibility.

The merits of this are the clarity of the patterns from a student's point of view, the proximity of the translations, ensuring language is 100% understood, and then the variety of ways the table can be used to generate more input and output. These include ways to generate communicative exchanges. I have blogged about that here.

Once again, the input in sentence builders is very simplified (how much depends on the level of detail provided). Most examples of sentence builders are not focused on cultural input - they are there to model samples of useful language matched to communicative goals. There is no reason why you shouldn't design sentence builders with culture in mind. This may be easier at intermediate level - for example designing sentence builders around a visit to a place, geographical information or a cultural visit.

Dialogues

In the audiolingual method the dialogue was the go-to means of presenting model language. A key difference with dialogues, is that they are written representations of spoken language so have the potential to offer a better model of authentic speech. In reality, written dialogues also feature somewhat stilted, simplified and unnatural language. (Just try asking ChatGPT to write a dialogue!)

As with written texts, new language can be glossed or parallel translations for the whole dialogue can be provided. The written dialogue also offers, as with the above two ways of modelling language, the opportunity to design sequences which allow for recycling of language. Words and chunks in a dialogue can be underlined, with alternatives provided after the text. Students can then slot in alternatives to the original dialogue.

Some teachers, with some classes, might prefer this arguably more authentic way of modelling language.

Knowledge Organisers

These, typically, are those booklets which contain sample sentences on the left, with parallel translations on the right. As with sentence builders, all the language is comprehensible and the KOs can be exploited in multiple ways, including slotting in new items into the model sentences. Repetitions are generated through, for example, translations, transcribing, dictation-translation, rote learning of sentences.

Unlike sentence builders the precise sentence structure is not indicated by columns. On the other hand, the sentences have the potential to contain a wider variety of structures.

I think critics of both sentence builders and knowledge organisers might argue that they are too limited in scope for the fastest learners and encourage too much rote learning at the expense of creativity. I am not sure this is true, since much depends on how you exploit them. Remember, they are a starting point in a sequence and that the language they model can be recycled in other forms later, some of which may be more creative and communicative.

Audio

Aural text is another way of modelling language, but clearly comes with much greater challenges. The fleeting nature of auditory input means it is much harder to process than written language. For this reason, audio is often used alongside written text. I have head it argued that students can only really focus on either the written or aural text, so using both together is superfluous. I much prefer the argument that hearing text while reading it builds skill with sound-spelling correspondences (phonics), along side both reading and listening skill. I also like the argument that seeing the written text acts as a scaffold for the essentially difficult task of listening, especially for students with weaker skills or less efficient working memories. Gianfranco calls this scripted listening.

One way to make aural input (scripted or not) easier and to generate more repetitions is to do narrow listening. This is the same in principle as narrow reading, with a handful of short listening paragraphs being provided, each one containing the same or similar language items. By the way, if you don't provide a transcript, you can always scaffold the listening through the nature of any supporting exercises. For example, questions in English will contains clues to the meaning.

As well as audio, we have the option of video, of course. Some would argue that it much more authentic to present aural language in this way.

Picture stories

I throw this into the mix, partly since it was a favorite of mine. This where you present a series of images which depict a series of events or short narrative. The input is provided by the teacher who narrates the story, building in pauses, repetitions and paraphrases. The pictures provide a scaffold for the interpretation of meaning, but the input still has the disadvantage of being unsupported by written text. For the reasons mentioned above, you may see this as an advantage for some classes, as they work hard to grasp meaning from the aural input and images. the teacher can provide 'aural glosses' from time to time (translations) for new or harder words.

The other advantage of this form of modelled input is that the pictures draw attention, perhaps even more so if they are striking or amusing.

Now, I am aware that input can be provided in other ways. For examples, teachers working within the TPRS paradigm like to co-create stories with classes with the aim, quite often, of recycling what they might call core language, many times over. the repetitions can be generated using circling (asking questions at various levels - yes/no, either/or, multi-choice and so on). In CLIL (Content and Language Integrated Learning) and PBL (Project-Based Learning) the input is much less tightly patterned and usually harder. In TBLT (Task-Based Language Learning), the input forms part of the task.

With all the above ways of providing input, what counts is not just the nature of the input, but how it is exploited. 

• Is the language always highly comprehensible? 
• Are there lots of opportunities to recycle, allowing students more chance to process and acquire the language longer term? 
• Do the associated activities allow students to rehearse language in working memory (giving it more chance of ending up in long-term memory where we want it)? 
• Are the activities enjoyable - stimulating, challenging enough or gamified? 
• Is the input inherently interesting - this is the largest challenge in a way; in the examples above the level of interest may vary, so we are stuck with the old language teacher conundrum that we want to provide simplified, patterned input, but we also want the meaning to be interesting. We generally tend to favour the first part! This is partly why some advocate the greater use of more complex and interesting texts where students have to work a bit harder to decode the meaning. Here is an example.
• Does the input and the activities generate student self-efficacy (Bandura) or competence (Deci and Ryan)? Is everything feasible and the right level of challenge?