Listening matters: Top-down and bottom-up listening

Introduction

This article, in a way, carries on from the one on Process listening. The terms top down and bottom up refer to ways in which someone might process a listening text. In many books on listening they are seen as two separate strategies that can be used by students to process and understand listening texts, but what exactly do they mean?

What do the terms mean?

Top-down strategies focus on the ‘big’ picture and general meaning of a listening text. Often the starting point is to discuss the topic and then to use a ‘gist’ or ‘extensive’ task to listen for the overall meaning. Top-down strategies rely on students knowing something about the topic and either knowing how particular exchanges in certain social situations work (i.e. the functional and situational language common to certain exchanges) or what ‘chunks’ of language (expressions, etc) ‘fit’ the particular topic or situation.

Bottom-up strategies, on the other hand, focus on listening for details and involve tasks that focus on understanding at a sound or word level. Tasks are ‘intensive’, as they focus on looking for particular details.

How do we process listening texts?

When we listen to our first language, or L1, we probably employ mostly top-down strategies, but for a second language, or L2, it is quite likely to be the other way around. However, as we will see later on in this article, it is quite difficult to separate the two when it comes to actually analyzing the strategies we have used on any one particular listening text. Not at a strategy level, as such, but at a time-sequence level – i.e. which strategy is employed first, and at which subsequent stage a different type of strategy kicks in.

Which strategy do students usually use?

It seems that in many cases students are locked into a bottom-up learning mode. By this I mean that they are desperately trying to understand every single word within a listening text. Unfortunately, in many situations this is just impossible. The main reason for this is the speed of ‘real’ talk. It is not that English is any faster than other languages, but simply that all languages are spoken at a speed where it is virtually impossible for the brain to process every word. Therefore, students often find listening very difficult. They become frustrated and demoralized and often give up.

However, this does not necessarily mean that the answer is to simply train our students to use top-down strategies. Although it might be helpful to start by understanding the overall gist of a listening text, it is not the final aim. Therefore, however much we might want to focus on top-down we cannot ignore bottom-up strategies either.

Is it an either/or situation?

No, not at all. To think that you can only process a listening from the general level (top-down) or in a word-by-word (bottom-up) way is very misleading. In fact it is quite possible, and I would argue quite common, to employ both approaches on any one text.

A good example of this might be when you are not really listening to something – it just happens to be on in the background, and you pick up on one or two words. Suddenly, you might be interested in the text, but of course you have not heard everything. So, what do you do? Well, you might try and reconstruct the sentence that you heard fragments of.

In order to do that you will use two strategies. You will take your knowledge of grammar and syntax to try and put together, piece by piece, a sentence that makes sense grammatically. This is a bottom-up approach. However, you might also take the words you heard and think about what they refer to. You will often use your knowledge of what those words might refer to on a content level and try to put together the likely sentence using your topic/subject knowledge. To all intents and purposes you are using a top-down approach.

My argument here is that you might well use a combination of the two, one feeding into the other and vice versa: a kind of loop effect, where it would be extremely difficult to say which came first.

Does it help to tell students not to understand every word?

The answer to that is ‘yes’ and ‘no’. Yes, because it might help some students relax and also realize that they aren’t expected to understand every word. But equally the answer might be 'no', because just telling students it isn’t necessary doesn’t really help them. It seems as though many students are reluctant to abandon attempts to understand every single word. What is probably more effective is using activities where it is impossible to get every word and yet the activity is doable. It is a careful balancing act, because you don’t want the student to simply switch off and not attempt the activity in the first place. This ‘consciousness-raising’ approach can be extremely effective and is often reinforced by discussing what actually happened after the task is complete.

It’s also important to remember that when we tell students that they only need to understand the key words, this advice is difficult to put into practice. How do the students know which of the words are ‘key’ ones? If they hear particular words, but not others, can they be sure the ones they heard were the key ones? So, when we give advice, however well intentioned it may be, we really need to be thinking about whether it is feasible to follow.

But can’t we just transfer L1 listening skills to L2 listening?

Although we have already looked at this in the article on process listening, I’d like to revisit it here as I think it is very pertinent. The first thing to say is that most people are unaware of exactly how they listen in their L1. Trying to talk about the strategies and sub-skills they employ might seem to be a good idea but may actually not move us any further forward. My argument here is that we are far better off concentrating on using practical activities designed to use certain skills and strategies and then discussing these in terms of how the tasks were tackled.

The second point is that trying to help students become as proficient at listening in English as they are in their L1 is probably futile. It may well be that how we listen and comprehend in an L2 differs – whether that's slightly or greatly – from how we listen in our L1. In fact, we may be (and become) far more conscious of how we listen in an L2 and how we process the information than we ever would in our L1. Instead of trying to transfer skills from L1 listening to L2 listening, we may be better off employing our time focusing on the skills and strategies needed. If they are ones that are already used and transferable, fantastic. If not, it doesn’t matter – we can still work on them if they are useful.

Are there any ways we can help our students with bottom-up strategies?

Yes, definitely. One of the biggest problems with a lot of listening texts seems to be length, and this is particularly true of higher levels (i.e. upper intermediate and advanced). It seems as though many writers, publishers and teachers seem to equate the length of a text with complexity and level. There is, in fact, an argument for having much shorter listening texts, some as short as one- or two-line dialogues, in order to be able to focus more on understanding sounds and words (especially in fast speech) and help our students process listening texts in a bottom-up way.

Some practical ideas

Listening bingo
Ask the students to draw a grid or table with six boxes (you can use more for higher levels – e.g. nine boxes at upper intermediate). Tell them you will play a piece of audio, and tell them the topic of the audio. (If you want, you can give a bit more information – e.g. You will hear two people talking about their plans for the weekend.) Ask the students to write a word or phrase in each box. These should be things they think they will hear during the audio. Monitor and check they have completed their grids. Play the audio. Every time a student hears a word or phrase in their grid they should cross it out. If they cross out all six they should put their hand up in the air (or shout, ‘Bingo!’).

Predicting vocabulary based on the topic is a skill that we all employ in our L1 both before and while we are listening. The task starts off with a top-down activity, predicting based on what we already know about the topic, but during the actual listening phase the focus is far more on a bottom-up process.

Dictogloss
Tell the students they are going to hear a short text (a few sentences or a short paragraph). Ask the students to put their pens down and just listen. Play the listening, or read it out once, and then ask the students to note down all the words they can remember – this should be done focusing on key words and NOT trying to remember everything verbatim. Play or read the text again, and then ask the students to work in pairs and reconstruct as much of the text as they can. Repeat the process one more time, and then pair the pairs and get them to compare their texts. Finally, compare their texts to the original and discuss.

Initially, this activity is bottom-up. However, as parts of the text are constructed, students will use the co-text (working out content and language based on what has already been said or what surrounds a particular utterance) to help build the rest of the text. This then moves from bottom-up to top-down strategies and often employs both simultaneously.

Gap-fill
Take a listening text and remove some of the words. Play the text and ask students to fill in the blanks.
Although this is a standard bottom-up approach it is amazing to see how much can be predicted and therefore how even the most bottom-up activities employ a degree of top-down processing.

Conclusion

As we can see from the activities above, it is almost impossible (and unnecessary) to separate listening tasks and strategies. It is far better to use ‘good’ listening activities and to talk about how students processed what they heard in order to do the task. If students are unable to do a particular task, then it is equally important to talk about why and see if it is possible to overcome the problems. The focus should not be on getting the right or wrong answers, but how these answers were reached.