Going SOLO - assessing undertanding of language concepts

Ever since my friend Darren Mead got me into SOLO taxonomy I have been cogitating and playing with different ways to use it in the learning of another language. That got me into thinking about what learning another language actually means. I know that it has nothing to do with national curriculum levels as they just vaguely describe some stuff that you can do (but obviously not that you would do in the real world!). I know it is definitely not GCSE controlled assessments which I believe they are going to make into a game show soon... ("So ladies and gentlemen you have to memorise 5 paragraphs and an exclamation mark in 6 hours otherwise you will fall into a pool of cold custard and exam markers! On your marks...") So I finally came to the conclusion that learning a language is about understanding and applying the rules that make up a language. Being a real linguist is about making connections which are then demonstrated in the language we produce and are able to understand. Of course there are loads of way to take in language and reproduce it (from flashcards to thinking skills) but it is not until we make concrete connections that we can really understand how language works and therefore become a linguist. 

So how can SOLO help? Well, SOLO (Structure of Observed Learning Outcomes [Biggs and Collis 1982])is a taxonomy which measures a learner's understanding of a concept by dividing understanding into five levels. 

The first level of understanding is called pre-structural. This is when a learner has no idea about a concept. So next time you are observed by OFSTED and the kids don't learn anything tell the Inspector that it was not unsatisfactory, simply pre-structural...

 
The next level is known as unistructural. This means that a learner has grasped one element of a concept. When Darren described this to me, he said that it was like having one brick but not really knowing how to make that brick into a house. In language learning terms, this can mean undestanding one aspect of a grammatical concept like gender and adjectival agreement- i.e. "I know what a noun is" or "I know that a noun is either masculine or feminine in French."

 

The next level is the multistructural level. We now have a pile of bricks, but we have not yet seen how they fit together to build a house. "I can identify the gender of nouns from the word that comes before it. I have more knowledge but I have not yet linked it all together."

 

This level is known as relational. "Finally, I can see that by fitting these bricks together, I can start to build a house. I know that nouns have a gender, I can identify the gender, I know that adjectives change depending on gender and I can now understand how all of this fits together, I can see the links between the gender of a noun and the way an adjective behaves. I have my house."

            
The final level is extended abstract. This means that learners can predict, using their          relational understanding, beyond the oricginal concept. They, for example, can guess that nouns and adjectives change when plural. They can start with the notion of gender as a unistructural part of a bigger understanding of the language they are learning. This is like seeing your house for the first time as part of a street or town.

So imagine applying this to the perfect tense. "I understand that tense is when we talk about different times. (unistructural). I know the verbs avoir and être. I know how to make a past participle. (multi structural) I know how to put together an auxiliary verb and a past participle. I understand that we need to apply my understanding of gender to verbs which take être. (relational) I wonder if I change the tense of the auxiliary verb that I could create the future, conditional or pluperfect...." (extended abstract - well it would be nice if they could get there!!!)

I am not entirely there with SOLO yet but I feel that I am at least multistructural so I will leave you with two parting gifts; firstly a link to a Gender & Solo document I have made with SOLO rubrics and ideas for activities and also a video from the brilliant www.hooked-on-thiniking.com with some gorgeous little Kiwi learners explaining their understanding of how the SOLO taxonomy has helped them!

 

Comments

[Gwyn]

You know Chris, I've a BIG problem with the SOLO taxonomy. It's those icons. Especially the last one with the funny thing coming out of the top, it reminds me of an Oli Bolly Dob Dob flower. I can't take it seriously. So unfortunately, everything you are doing is fundamentally flawed. You can tell Darren too.

[Steve]

My problem with this is that it works poorly as a model for language learning. It's OK on the conscious learning level, but real competency emerges naturally through a process of natural acquisition, as in first language learning. Most kids can work out how a grammar rule works, but they have little or no fluency. Knowing a grammar rule, but not being able to really internalise it is pretty useless.

[Chris Harte]

Hi Steve, I have not been clear about the purpose of the SOLO taxonomy, it is not a model of language learning, it is simply a way of measuring linguistic understanding, an effective tool which looks at the explicit connections children make to understand a concept, in this case a grammatical concept. I would question your assertion about competency emerges through a process of natural acquisition - in a school context, we strive to give students as much exposure to authentic or semi authentic text/listening materials (something you do wonderfully and share on www.frenchteacher.net) but to arrive at linguistic competency through natural acquisition in this context is nigh on impossible due to the lack of time. I would argue that the "conscious learning" which SOLO measures is exactly what we need to be doing to support "natural acquisition". I suppose the analogy us that my son is bilingual as my French wife and I only speak French at home. He is 5 and speaks pretty fluent French and English but had no concept of how gender and adjectival endings work, he just does them. This had taken 5 years of immersion. I hope (as Daddys do) that he will become a good linguist with the capacity to understand the grammatical structure of any language he chooses to learn. Being able to see the links he has explicitly made in terms of the language structures he knows will help me to guide him towards fully understanding the grammar concept and applying it. Indeed I would argue that talking explicitly about language understanding is the best way to internarlise a grammatical concept. I hope I have made mypoint more clearly and not muddied the waters!
Cheers Chris.

[Darren]

Hi Steve

Have you tried to build a teaching and learning model around Solo? Even a failed one would be interesting to hear about?

I also disagree about knowing a grammar rule. As someone who only genuinely started to learn a second language in my twenties knowing a rule is a strategy I use. Particularly when reading. SOLO taxonomy facilitates the metacognition necessary to learn from experience and the associated mistakes.

Cheers

Darren

[Artichoke]

Thanks for this post Chris – I love what you are thinking about with respect to a model of learning outcomes and how it might be valuable to students – your post identifies some parallels between language acquisition and mathematics. I have pasted your examples into a mathematical data rubric so you will understand what I am getting at.

Extended abstract
The use of data elements external to the system is a feature of an extended abstract response. The generalisation of the elements takes account of new and more abstract features.
"I understand that tense is when we talk about different times. (unistructural). I know the verbs avoir and être. I know how to make a past participle. (multi structural) I know how to put together an auxiliary verb and a past participle. I understand that we need to apply my understanding of gender to verbs which take être. (relational) I wonder if I change the tense of the auxiliary verb that I could create the future, conditional or pluperfect...." (extended abstract)
Relational
A relational response reflects the ability to integrate the elements and operations of the question in a way that enables an overview of the stimulus item. Children using an algorithm at this level would be able to check for errors and inconsistencies, and would be able to reconstruct missing elements of the algorithm. Features of responses at this level include the ability to reverse operations and the set of elements used are internal to the system
"I understand that tense is when we talk about different times. (unistructural). I know the verbs avoir and être. I know how to make a past participle. (multi structural) I know how to put together an auxiliary verb and a past participle. I understand that we need to apply my understanding of gender to verbs which take être. (relational)
Multistructural
The learner at this level can use multiple data elements, but the elements are not integrated. Hence the response can consist of a number of discrete closures. Typical of these responses would be following strict algorithmic procedures that involve a number of steps. However, if a single step was forgotten, or an error made, the respondent would be unable to reconstruct the algorithm. This lack of an overview of the data elements and their relationships makes the response patterns inherently unstable and thus considerable variability may be expected from children responding at this level.
"I understand that tense is when we talk about different times. (unistructural). I know the verbs avoir and être. I know how to make a past participle. (multi structural)
Unistructural
This set of responses uses only one relevant element of data from the stimulus item. A feature of responses at this level is the desire to close quickly and to ignore inconsistencies that may result from the response.
"I understand that tense is when we talk about different times. (unistructural).
Prestructural
The response indicates an inability to engage with the language task in a meaningful way. Such a response may involve restating the question, or focusing on some irrelevant data that is incidental to the question. It may reflect that the student is incapable of responding, or does not wish to respond, in the target mode.

The issue that Steve raises wrt language acquisition is also relevant in discussion around teaching and learning in mathematics – they talk about how instrumental instruction influences relational learning understanding – and that is why you have to be cautious with rubrics that look like they could be used for instrumental instruction / teaching progression (as happened with the mathematics progressions in the numeracy project in NZ have been) - this defeats the whole purpose of the rubric to help students self-assess their own learning outcomes. .
The way to escape this is to tip it on its head and look at it from a solving a problem perspective - unknown problem maths/ descriptive writing in languages. Set a writing challenge – e.g. family holiday - frame it in a way that to complete the students will have to use the language structures appropriate for the level they are working at. Then code the outcomes for this writing activity as success criteria against SOLO outcomes. This allows you to escape that reductionism - lock step – do this – now do this - which if used as a teaching method instead of a way of looking at learning outcomes - students will simply memorise the progressions and follow them instrumentally without understanding relationally – the why stuff.

You will need to start with a situation in languages that provides opportunities to apply knowledge or understanding of the language structures, grammar, tenses etc for the level your students are working at. Say – writing about a family holiday If they show in their writing about [a family holiday] that they have used a range of structures, appropriate vocabulary, to construct sentences about the holiday etc the outcome is multistructural. If they go further and construct sentences that follow a logical sequence, or connecting different representations together to explain, compare or relate the ideas to the [family holiday] then is more like a relational outcome. If they do all of this correctly and tie it together to forms a generalisation, personal impression, insight or prediction about the next family holiday etc (realise they will not have the tenses to allow some of this in early stages) then they are showing an extended abstract learning outcome.

I can send you some research readings around mathematics and some languages rubrics NZ teachers are working on if you think it will help.