All the courses mentioned should be so constructed that the students can learn from the printed or tape-recorded material without the

------------------------------------------------- page 17 --------------------------------------------------

permanent supervision of the teacher, and the need for students to have to call for the teacher's help should be reduced as much as possible in gradually improved versions of each programme.

The role of the teacher

What work is left for the teacher? He has to supervise students working on programmes to ensure that they reach the required standard of performance before proceeding from one unit to the next. After some time, he will find out which students can be trusted when doing what, and he can concentrate his supervision on those few students who cannot be trusted. He has to decide for each student which courses he should work through and at which point in the total teaching programme the student should start on a new course. He has to decide the proportion of time a student should spend on any particular course. This depends on the ability of the student for the particular skill taught in a certain course and also on the learning objectives of the student. The teacher administers and evaluates regular tests. On the basis of the test results, he may change the overall teaching programme for a student.

When the teacher is satisfied that all students

------------------------------------------------- page 18 --------------------------------------------------

are busy, he can, in a special section of the language laboratory, or rather in a room next to it, conduct those parts of language learning which cannot be, or have not yet been, programmed. (FN 5) The one thing he will not do is introduce new material, new structures, phrases, words, etc., to his students. All that can be programmed. The teacher will conduct conversation practice, with groups of not more than four students, once these students have mastered the basic structure of the language through the use of programmes. He will take part in intensive conversation exercises with one student at a time, the conversations and the teacher's corrections and suggestions for exercises being recorded and later analysed and followed up by the student. The teacher will analyse conversation tapes that have been produced by two students without his taking part. The students will be present while the 'marking' of the tape takes place, and their queries and the teacher's comments will be recorded. (Recording of the teacher's comments always saves time!) The teacher will discuss essays and other written work with the students concerned. As a matter of principle, whatever work is done is only done with those students who are directly concerned. In conventional classes students lose too much valuable time by having to listen to corrections and discussions of mistakes which

------------------------------------------------- page 19 --------------------------------------------------

they themselves have not made. But backed by programmed courses, the teacher is able to give 'private lessons' to his students, unless the subject matter demands group work as a positive feature rather than an economical necessity, e.g. in small conversation groups where we have not only the direct inter-action between teacher and student but also a positive inter-action between student and student.

Language Gymnasium

I am at present * running an experimental German class at Woolwich Polytechnic which seems to indicate that such a plan is workable. It is workable despite the fact that we at Woolwich have as yet only very few of the programmes that I have described as desirable. Students were told at the beginning of the course that I was not going to teach them but that this was a self-study group and they had to do all the work. I was ready to assist them whenever asked for help but they had to do the pushing. And they are doing the pushing. I have never had such a lively class before. Students come to remind me that they want to do more audio-visual work (this is done live as a sub-programme) because they haven't done it for some time, that they want to

Footnote: * i.e. February 1965, when this paper was first presented to the London Branch of the Association for Programmed Learning.

------------------------------------------------- page 20 --------------------------------------------------

do conversation again, that they want special tapes for listening practice (they can get the programmed tapes for themselves). These reminders are necessary because in a class in which every student does something different it is impossible for the teacher to remember each student's requirements in detail. Students are told at the beginning of their course what type of work they are supposed to do, for what purpose and in what proportion, and it is up to them to make sure they get it. Expressly handing the responsibility for their progress over to the students seems to have a good motivational effect.

In the Woolwich class, we have students who are absolute beginners and others who have an experience of German equivalent to 'A' level but not the required standard of accuracy. The more advanced the students are, the less use of programmes can be made and the more teacher time they need. Therefore in one class a large number of beginners but only few advanced students can be accommodated. The beginners will hardly require any attention and leave the teacher free to deal with the advanced students. It should be easy enough to fill classes in this way because there are normally more beginners of a language than there are advanced students. But even if a class starts with beginners only,

------------------------------------------------- page 21 --------------------------------------------------

after some time the better learners will get ahead of the rest and therefore be able to demand and get more teacher time.

Among the courses at present available at Woolwich is a programme 'Introduction to German' written by myself, the Temac German programme *, remedial tape exercises written by myself, and other non-programmed practice tapes produced by commercial firms. My own Introduction is in its first draft stage, does its job quite satisfactorily but will, I hope, be scrapped by the end of the year. (FN 6) It provides about one term's work for an average student coming to school one evening a week for a three-hour session minus a long break, and not doing any homework, i.e. a total of about 25 hours' learning time. In this course, the student learns a vocabulary of about 160 words, all nouns being masculine; the nominative case of nouns only; the plurals of all nouns occurring; including some irregular verbs; some of the socalled separable verbs in contrasting German and English sentences; numbers up to thirty (they are not included in the above vocabulary count); the names of the letters of the alphabet (so that he can spell his name over the phone, for instance); the relationships between German sounds and spelling and dictation of all words and sentences

* Published by Encyclopaedia Britannica Ltd.

------------------------------------------------- page 22 --------------------------------------------------

occurring. The materials consist of a tape, notes explaining points of grammar and learning procedure and giving the solutions for the spelling exercises, a set of drawings which serve as stimuli for the tape-recorded exercises, and the recording scripts, which are given to the students only when they have completed the programme.

Types of language programme

In the past, two basic approaches to language programming have become known. Professor Rand Morton's fascinating techniques described in his paper 'The language laboratory as a teaching machine', have met with a great deal of resistance in this country because they sidestep meaning for such a long time. (They do so for good reason.) Apart from this, Temac language programmes have been introduced to this country from America but are only sold on a limited scale. (It is now being investigated to what extent these programmes are suitable for the British educational system and in what respects they have to be modified.) Externally they closely resemble the linear programmes used in subjects outside the language field. Their principle is that the student should (like in most conventional linear programmes) not make any mistakes and should be so prompted that he gets

------------------------------------------------- page 23 --------------------------------------------------

his answers right at first attempt. This procedure can be very boring for good students, particularly as in languages intelligent prompts can hardly ever be given. I therefore think that the conventional linear programme is not very suitable for language learning* and have developed the 'Repetitive-unit programme', a programme pattern specially aimed at language learning with today's ordinary tape-recorders, in the language laboratory or at home. The system is described here as a frame-work into which language teachers can fit their own materials for the language laboratory.

RU-programmes

All items are presented in units of ten. Each item consists of a stimulus and its response. As a rule, all the items in a unit demonstrate, or demand application of, the same 'rule' (the 'rule' need not be expressly stated), that is, they all have the same structure and the same stimulus / response relationship. Each item has a

Footnote: *This has recently been confirmed by the publication of A K Tyrer's 'A programmed German grammar', Parts I and II, Methuen, London, 1965 and 1966 respectively, a book that has done much to discredit programmed learning among language teachers. The programme adheres slavishly to linear programming principles, aspires no higher than to teach paradigms and isolates grammar from language usage. Although programmed learning techniques can be misused in this way, they should not be identified with bad and outdated teaching aims and methods. In fact, the best that audio-visual and audio-lingual techniques have to offer can and should be utilised in the framework of programmed learning.

------------------------------------------------- page 24 --------------------------------------------------

number which is called out before the stimulus. The student has a printed scoring book (see sample page below) which has numbers corresponding to the item numbers in a vertical column. Depending on the teacher's preference, the student either listens to each unit once or twice before trying to respond or he tries to respond at first presentation even though the unit may contain entirely fresh material. This first attempt at responding is called 'test run'. The student ticks off all items he has answered correctly. He counts the correct responses at the end of each run through the unit and enters the total at the bottom of the current column. If during a test run he has made ten correct responses, i.e. 100%, he proceeds to the next unit.

Intensive runs

If the student makes too many mistakes or if his mistakes occur in a certain pattern, the principle of 'intensive runs' comes into operation. An 'intensive run' is concentrated practice of one item or a small number of consecutive items. Intensive runs are marked in the scoring book by a dot for each correct response. No blanks are left for wrong responses. The same column is used again and again until the next through run is due.

------------------------------------------------- page 25 --------------------------------------------------

An intensive run tackles a group of three or four items at a time. Work on the group continues until at least three correct responses (consecutively or not) have been given for each critical item. The critical items are those for whose sake the intensive run is being done. For instance, if items 2 and 3 call for an intensive run, a non-critical item (either item 1 or item 4) is taken into the group to make up the minimum number of three items for a group.

This minimum number ensures that speaking does not become mechanical, that the student has a slight chance of forgetting a critical item and there is always a genuine task before him. The maximum number of items in a group ensures that the critical items are presented to the student with a minimum frequency. As the student is likely to get the non-critical items right at the first attempt at the group, but the critical items only after several attempts, the scoring book often shows more than the minimum number of three dots placed against a non-critical item whereas a critical item often receives no more than the minimum of three correct responses.

When the intensive work on one group is completed, the student proceeds to the next group (still using the same scoring column), and so on until all groups containing critical items

------------------------------------------------- page 26 --------------------------------------------------

have been done. Then the next through run is due, which will show the results of the intensive work.

'Groups' in an intensive run

If, in the test run, the student counts three or fewer correct responses, he has to make an intensive run. In this case, he works through the following 'groups' of the unit in turn: Group A: Items 1 to 3, Group B: Items 4 to 6, Group C: Items 7 to 10. After that, the next through run is due (marked in a fresh column and using ticks instead of dots) and the progress is scored. The student continues working on the unit until he has reached the passing score, defined as 'two consecutive runs with no errors'.

If the score shows two or more consecutive items wrong in the same run (= vertically consecutive blanks in the scoring book),

------------------------------------------------- page 27 --------------------------------------------------

the student does an intensive run covering these critical items, including, if necessary, a non-critical item. The reason for this rule is that individual items are comparatively difficult to locate on tape and consecutive errors are therefore a conveniently large target for attack.

If the score shows the same item wrong in two consecutive runs (= horizontally consecutive blanks in the scoring book),

the student does an intensive run covering that one critical item. To make up a group, he includes two neighbouring, non-critical items. Although it appears to be taking a lot of trouble to wind the tape back to locate and practise just one item, this labour is justified because this item has shown itself to be difficult for the student in question and the student has during two 'through runs' made no progress with it. Further through runs are therefore not justified but more stringent measures (namely intensive runs) must be taken.

------------------------------------------------- page 28 --------------------------------------------------

Scoring machine

In the early years of programmed learning, there existed a simple linear teaching machine, known as the 'Skinner disk' . (FN 7) The programme was printed on a circular piece of card-board with a capacity of 30 items. The student decided, by comparing with a model answer, if his own answer had been correct and accordingly pushed either a 'Right' button or a 'Wrong' button to inform the machine of his performance. As the disk turned, the machine kept presenting those items to the student which he had answered wrongly while dropping those items which he had answered correctly. The machine could be set in such a way that the student had to answer a question correctly n consecutive times before the machine would drop the question. In the language learning system described above, the student takes over functions which the 'Skinner disk' fulfilled automatically but which today's usual tape-recorders do not perform.

The reader may think that the RU scoring system is so involved that the students spend more energy on trying to remember and apply the scoring rules than on learning the foreign language. This is in fact not the case:

1. The RU system in practice is not as complicated as it appears to be when all

------------------------------------------------- page 29 --------------------------------------------------

the rules and their justification are compressed into very little space.

2. The time spent on learning the RU system is saved over and over again by the increased speed of language learning later on.

3. At present, I spend about three hours at the beginning of each course teaching the students the scoring system and after that monitor not only their linguistic performance but also their working method. I have somewhat reduced this time by handing out a straightforward description of the RU system for study at home before starting the course and hope to save further time by explaining the method in a linear programme.

4. I have used the RU system over several years and most students, both the highly intelligent and the rear-guard, appreciate the system and have no difficulty in getting accustomed to it.

5. I hope that one day machines will be introduced into schools capable of assessing the students' performance. The RU system, with some amendments, would provide a basis on which the machines can decide in which sequence to present and revise the items.

------------------------------------------------- page 30 --------------------------------------------------

RU algorithms

The following algorithms state with greater precision the learning procedures described above. They indicate how a machine could control this type of work, although the machine algorithm would not necessarily follow the same principles as the algorithm suggested here. Each of the higher algorithms presupposes that there are sub-algorithms specifying the meaning of the complex operations in the higher algorithm, until we come down to a sub.-algorithm all whose operators represent only elementary steps. The sub-algorithm 'Do item j' (below) comes quite close to such a final sub-algorithm. It does however contain an operator for which I have no algorithm at present (if one can ever be found at all). This is the operator 'Compare your own response and the model response'. The problem hinges on the definition of the term 'identical'. The interpretation of this term constitutes not only a major problem in general linguistics (cf. Twaddell, 1935) but is also one of the grounds for the assertion that languages cannot be learnt without the help of the (evaluating) teacher. The difficulty of constructing an algorithm thus pinpoints a major teaching and programming problem. In order to reduce the ill effects of the absence of algorithms for comparing two utterances and deciding whether or

------------------------------------------------- page 31 --------------------------------------------------

not they are the same, I have tried, in my paper 'Problems of task analysis for language programming', to confine the problem to section 2 (Imitative articulation) of the task diagram (where the teacher monitors) instead of allowing it to infest a whole language course, which is the case in much language laboratory work today.

RU main algorithm

The underlying structure of the RU main algorithm as presented below is the following:

1. Do through run

2. Check on 10 items correct

3. Check on 3 or less items correct

4. Check on convenient target for attack

a. 3 or more consecutive blanks

b. 2 consecutive blanks

5. Check on difficult items

a. Single blanks horizontally adjacent

6. Ensure two consecutive through runs 100% correct

All the student's or machine's actions are denoted by capital letters. The conditions for certain actions are denoted by small letters. Presence or absence of a condition is denoted by plus and minus signs respectively.

------------------------------------------------- page 32 --------------------------------------------------

In the RU main algorithm, 'column j' is the column containing the last through run; 'column i' is the adjacent column on the left of column j. Column i may be a 'through run column' or an 'intensive run column'.

Logical conditions

a Are 10 items correct in column j?

b Are 10 items correct in column i?

c Are 3 or less items correct in column j?

d Are there three or more vertically consecutive blanks in column j?

e Are there two vertically consecutive blanks in column j?

f Are there any single blanks in column j?

g Are there three or more vertically consecutive blanks further down in column j?

h Are there two vertically consecutive blanks further down in column j?

i Are there any single blanks further down in column j?

j Are the first three of these blanks untreated by an intensive run?

k Are the first two of these blanks untreated by an intensive run?

l Is the first-of these single blanks untreated by an intensive run?

m Are these two blank items items 9 and 10?

n Is this untreated blank item item 9 or 10?

------------------------------------------------- page 33 --------------------------------------------------

o Are any of these single blanks in column j adjacent to a blank in column i?

p Does column i contain a through run?

Operators

A Start

B Do a through run

C Do intensive run of items 1 to 3

D Do intensive run of items 3 to 6

E Do intensive run of items 7 to 10

F Do intensive run of items 8 to 10

G Do intensive run of these three consecutive blanks

H Do intensive run of the two blank items plus the item following them

I Do intensive run of the blank item plus the two items immediately preceding it

J Do intensive run of the blank item plus the two items immediately following it

K. Stop, or proceed to next unit

------------------------------------------------- page 34 --------------------------------------------------

------------------------------------------------- page 35 --------------------------------------------------

Operators C to J are identical ('Do intensive run') except that they specify different 'groups'. Their sub--algorithm is given below. The sub--algorithm for operator B is also given below. Operators A and K. represent elementary operations.

RU sub-algorithm 'Do a through run'

Logical conditions

NONE

Operators

A Start
B Move one column to the right
C Do item 1
D Do item 2
E Do item 3

... etc

L Do item 10
M. Stop

Figure 2: RU sub-algorithm 'Do a through run'

------------------------------------------------- page 36 --------------------------------------------------

In this sub--algorithm, operators A, B and M. are elementary. Operators C to L all have the same sub-algorithm 'Do item j', given below.

RU sub-algorithm 'Do an intensive run of 3 (or 4) items

Logical conditions

a Is there an item d?
b Is a a critical item?
c Have you scored three or more correct responses to the item in question?
d Is there a next item in the group?
e Is the next item a critical item?

Operators

A Start
B Do item a
C Do item b
D Do item c
E Do item d
F. Stop

------------------------------------------------- page 37 --------------------------------------------------

Figure 3: RU sub-algorithm 'Do an intensive run of 3 or 4 items'

------------------------------------------------- page 38 --------------------------------------------------

In this algorithm, A and F. represent elementary operations. B and E have a sub-algorithm 'Do item j', which is given below.

RU sub-algorithm 'Do item j'

Logical conditions

a Were your own response and the model response identical?

Operators

A Start
B Present stimulus j
C Listen to stimulus j
D Respond to stimulus j
E Present model response j
F Listen to model response j
G Compare your own response and the model response
H Score 'Correct'
I Leave blank
J Repeat model response
K. Stop

Figure 4: Sub-algoriothm 'Do item j'

------------------------------------------------- page 39 --------------------------------------------------

In this sub-algorithm, A, C, F, H, I, J and K. are elementary operations. B and E are elementary in the terms and at the level of RU algorithms (dealing with the selection of stimuli from a given set). The pushing of one or several buttons, moving of levers etc. (implied in B and E) can be regarded as elementary at this level of discussion. D requires the application of a linguistic algorithm (unless the task is simply the imitation of the stimulus), such as those suggested in my paper 'Problems of task analysis for language programming'. Since the present level of discussion is not linguistic, D too can be regarded as an elementary operation.

G can be algorithmised up to a certain extent. The student can, for instance, be instructed to compare each morpheme in turn and to check whether the morphemes are identical and whether they occur in the same order. Then, however, a further sub-algorithm would have to be created specifying when morphemes (in their acoustic representation) are and when they are not identical. This will get entangled in the problems of phonemics and phonetics. How can accuracy in imitation be measured and specified? We can only state the problem here. For the time being, the teacher and his intuition are our only refuge.

Insert: March 2011-03-17

Prose version of Main Algorithm (RULA, later named LASPEX)

The following prose version of this algorithm was added to this article in March 2011. It is identical with the diagram on p 34, but whereas the original article was written for a technically inclined readership, interested in the theory underlying the procedure and in its cogency, I have now added two prose versions to make it more accessible for language learners and teachers in general.

1 A: Start here.
· Go to B = 2.

2 B: Do a through run (i.e. try all items from 1 to 10)
· Go to a = 3.

3 a: Are 10 items correct in column j (the current column)?
· If yes, go to K. = 30 = Stop.
· If no, go to c = 4.

This step enforces the following rule: If during the test run (the first run through the exercise) the learner produces 10 correct responses, it is evidence that he knows the exercise so well that he can stop at this point.

If he has 9 or fewer correct responses during the test run, he must continue practising until he gets 10 responses correct TWICE RUNNING. This is the only way to make sure that he really knows the exercise well enough and retains it well enough in the long term. This is the zero-tolerance approach. It does not permit any uncertainties at any stage and thus can guarantee long-term mastery and total confidence.

4 c: Are 3 or fewer items correct in column j?
· If yes, go to C = 11.
· If no, go to d = 5.

This question establishes whether the whole of the exercise has to be treated by .intensive runs. (which is required when the learner has hardly any prior knowledge), or whether we can limit the intensive runs to certain unknown items in the exercise. Example: Column 1 on the scoring sheet.

5 d: Are there three or more vertically consecutive blanks in column j?
· If yes, go to j = 15.
· If no, go to e = 6.

Example: Column 1 on the scoring sheet.

6 e: Are there two vertically consecutive blanks in column j?
· If yes, go to k = 18.
· If no, go to f = 7.

7 f: Are there any single blanks in column j?
· If yes, go to p = 23.
· If no, go to B'. = 8.

8 B': Do a through run (i.e. try all items from 1 to 10)
· Go to a = 9.

The purpose of this operation is to measure progress made as a result of the intensive runs.

9 a: Are 10 items correct in column j (the current column)?
· If yes, go to b = 10.
· If no, go to c = 4.

10 b: Are 10 items correct in column i (the column left of column j, i.e. left of the current column)?
· If yes, go to K. = 30.
· If no, go to B' = 8.

We are checking here whether the learner has 10 items correct twice running. That is the target standard. When he has reached this, he can stop learning or move on to the next exercise. If he has not reached this standard, he has to continue working until he gets 10 items correct TWICE running.

11 C: Do intensive run of items 1 to 3.
· Go to D = 12.

In an intensive run you tackle a small group of adjacent items, in this case items 1 to 3, again and again until you have given at least three correct responses of each item. You mark each correct response with a dot. You do not mark wrong responses. Example: Column 2 on the scoring sheet.

12 D: Do intensive run of items 3 to 6.
· Go to E = 13.

13 E: Do intensive run of items 7 to 10.
· Go to B' ' = 14.

14 B: Do a through run.
· Go to c = 4.

A through run is more difficult than an intensive run and requires more skill because the items are more widely spaced and can only be answered correctly if you have acquired a longer retention span. This through run measures how much you have achieved during the preceding intensive runs. It is an indicator of your progress. It also determines what you have to do next on your route to mastery.

15 j: Are the first three of these blanks untreated by an intensive run?
· If yes, go to G = 16.
· If no, go to g = 17.

This question deals with the situation in which there are more than 3 consecutive blanks (because in most cases we want no more than 3 items in an intensive run) and ensures that the first three of these blanks are tackled first.

16 G: Do intensive run of these three consecutive blanks
· Go to g = 17.

17 g: Are there three or more vertically consecutive blanks further down in column j?
· If yes, go back to j = 15.
· If no, go to e = 6.
If there is another group of 3 consecutive blanks, it will now be eliminated. If there are only 2 consecutive blanks, they will be eliminated by another step.

This question ensures that all consecutive blanks in an exercise are gradually eliminated.

18 k: Are the first two of these blanks untreated by an intensive run?
· If yes, go to m = 19.
· If no, go to h = 20.

All pairs of consecutive blanks will now be eliminated one by one.

19 m: Are these two blanks items 9 and 10?
· If yes, go to F = 21.
· If no, go to H = 22.

If there are only two consecutive blanks (only two consecutive critical items), a non-critical item has to be added to form a group. A group consists of at least 3 items. We need three items in a group in order to increase the retention time required between repetitions. If the retention time is too short, learning is less effective and correct responses do not provide evidence of sufficient retention. Step 19 helps to determine unambiguously which non-critical item will be added to form the group.

20 h: Are there two vertically consecutive blanks further down in column j?
· If yes, go to k = 18.
· If no, go to f = 7.

21 F: Do intensive run of items 8 to 10.
· Go to f = 7.

22 H: Do intensive run of the two blank items plus the item following them.
· Go to h = 20.

23 p: Does column i contain a through run? (Column i is the column left of the current column.)
· If yes, go to o = 24.
· If no, go to B' = 8.

24 o: Are any of these single blanks in column j adjacent to a blank in column i?
· If yes, go to l = 25. (l = small l as in "lottery")
· If no, go to B' = 8.

This question establishes whether there are two horizontally adjacent items (indicator for a difficult item, needing intensive treatment, to avoid endless repetition of the whole exercise without any progress).

25 l: Is the first of these single blanks untreated by an intensive run?
(l = small l as in "lottery").
· If yes, go to n = 26.
· If no, go to i = 27.

26 n: Is this untreated blank item 9 or 10?
· If yes, go to I = 28. (I = capital I as in India).
· If no, go to J = 29.

27 i: Are there any single blanks further down in column j?
· If yes, go to o = 24.
· If no, go to B' = 8.

28 I: Do intensive run of the blank item plus the two items immediately preceding it. (I = capital I as in India)
· Go to B' = 8.

29 J: Do intensive run of the blank item plus the two items immediately following it.
· Go to i = 27.

30 K.: Stop, or proceed to next exercise.

Informal description of LASPEX-Preferred

Look at the scoring sheet which illustrates all situations which can possibly occur while working through a spoken IDYLL exercise.

Column 1:

You start with a test run to establish your prior knowledge of the exercise. In a test run you tackle one item after another, from 1 to 10. You make a tick for each correct answer and leave a blank for an incorrect answer.

Count the number of ticks (correct answers) and write it into the bottom row (Totals correct).

If you have 10 items correct in that column, you proceed to the next exercise.
If you have fewer than 10 items correct, you must continue practising until you have 10 items correct TWICE RUNNING (see the Totals in column 11 to 14). Column 11 was not sufficient proof of mastery. Column 12 proved that the good results in Column 11 were not mastery but luck.

If you have three items or fewer correct in Column 1, you must make several 'intensive runs'. Intensive runs enable you to focus on difficult items. Therefore they lead very quickly to mastery.

You divide the items of Column 1 into groups of 3 or 4 items: 1-3, 3-6. 7-10.

In intensive runs. you use the same column repeatedly and you mark each correct answer with a dot and not with a tick. You do not mark wrong answers.

You do each item in the group (e.g. 1-3) until you have at least 3 dots in each cell. That is the target for a group. Then you go to the next group, until you have mastered all groups in that column. You do NOT count and write down the total number of ticks in an intensive run. Example: Column 2.

Column 3

After each intensive run you do a "through run". Its purpose is to establish and record how much you have benefited from the intensive run. You expect the totals to increase.

At the end of each through run, e.g. Column 3, you write down the total. Then you inspect the column and look for two or more vertically adjacent blanks. There are two such blanks in Column 3. There are a "convenient target for attack" because they are adjacent, both are unknown, and you focus on them to eliminate them from the list of unknown items. You spin the cassette recording back to that position, pencil the position number into the cell for the first item, and then start practising that group.

Each group, however, requires three items to ensure that the intervals between repetitions of each item are long enough to give you a chance of forgetting and valid proof of success when you get them right. The items with blanks (where you gave the wrong answer) are called "critical items". Since in Column 3 you have only two critical items, you add a non-critical item (item 4) to make the group complete.

Column 4

You now do an intensive run on the items you have identified as critical in Column 3.

Column 5

Now you have to assess (measure) the results of the intensive run. You want to see the totals going up, and you want to see that you are getting the critical items right.

The total in Column 5 has gone up, and you got the critical item right.

You inspect column 5 and look for vertically adjacent blanks (convenient targets for attack), There are none. You have not yet reached the target standard [10,10]. Therefore you have to do another through run: Column 6.

Column 6

You compare the totals in column 5 and 6: no progress. If you did more through runs, you could carry on like this forever. You do not want to waste time. You have to concentrate on the two items which did not improve. You have to shorten the intervals between the repetitions of these items to anchor them in your memory. These are defined as "difficult items", i.e. items with vertically adjacent blanks. Difficult items are items which can not be mastered by "soft methods", i.e. by through runs.

Whenever you have two through runs in succession, you inspect the last two columns for vertically adjacent blanks. Item 2 is a 'difficult item' but to tackle it in an intensive run, you need a group of 3 items. You therefore add two non-critical items and get the group 1, 2, 3. Similarly you establish the group 8, 9, 10 at the end of the exercise.

Column 7

You do the intensive runs defined in the previous two columns.

Column 8

You measure progress by doing another through run. There is progress: 9 items correct. You cannot find a critical item in Column 8, therefore you do another through run.

Column 9

The total correct is 8, as opposed to 9 in the preceding column, i.e. your performance declines. That is a warning sign. You inspect both columns to identify the problem and then to tackle it with vigour.

You notice that items 5 and 6 are now wrong whereas they were right before. Don't be dismayed by that observation (as many students are). This is quite normal in human memory. What you know fades in and fades out. Observe the workings of your own mind with interest and slight amusement, never never never get angry with yourself, never say that you are a bad learner or have a bad memory. These comings and goings in memory are normal like the wind in nature blowing this way and that.

Just do what the algorithm tells you and gradually your memory will get steady and slightly more reliable, but never completely reliable.

You now form a group for an intensive run.

Column 10

Item 5 is defined as a 'difficult item' (horizontally adjacent blanks), and item 6 together with item 5 makes a 'convenient target for attack' (vertically adjacent blanks). Item 7 is a non-critical item used as padding to make the group complete.

Column 11

You check your progress by making a through run. Hurray, 10 items correct. However, this is not the end of the story. We must make sure that this is not an accidental success. The IDYLL METHOD gives you fantastic guarantees of success. This can only be achieved if you work like a master craftsman. Do everything to perfection. Follow the instructions to the letter. Do not take even the slightest risk. Your memory (which houses your knowledge) is like the vaults in the Bank of England: you do not want its treasures to be stolen.

You target of mastery is: 10 items correct TWICE RUNNING. Remember, it has happened before in this exercise that in one column you got an item right, and in a later column you got it wrong. That must not be tolerated.

So you do another through run.

Column 12

Lo and behold, your scepticism was justified. One item is wrong again. You have not 'forgotten' it. Nobody can say what you have forgotten or not, but you 'got it wrong', and that's what we go by. We don't care whether or not you 'forgot' something, but we give you procedures through which you can make sure that you 'get things right'.

You inspect Column 12 and look for a 'convenient target for attack' (vertically adjacent blanks). You cannot find any. Therefore you do another through run: 10 correct - and another through run: 10 correct.

Column 13 and 14

You have achieved mastery and can have a rest or move on to the next exercises.

LEARNING EFFICIENTLY IS THAT EASY. IT IS THAT FOOLPROOF.

('And God saw that it was good': the seventh day.

end of insert made on 2011-03-17

------------------------------------------------- page 40 --------------------------------------------------

Advantages of the RU system

The learning method influences the actual programme written for it and I suggest that there are, as far as language programmes are concerned, the following advantages over the linear method with small frames:

1. Since correct responses at first attempt are not expected, the programme can be more concise, and more challenging for the intelligent student. Each student hears one item only as often as he needs to hear it. This makes the RU programmes particularly suitable for brushing-up work.

2. As the units are small and an unlimited number of repetitions is possible, weak students are not discouraged.

3. Concentration is aided by the aim of mastering each unit with as few 'through runs' as possible. The system of intensive runs makes sure that the student does not go to sleep and repeats the whole unit indefinitely without ever coming to grips with the 'difficult' items. The student is never in doubt what to do next. The challenge is always before him.

4. Mastering one unit by reaching the passing score gives the student a feeling of well-deserved success. In a conventional linear programme, the mastery of one frame occurs

------------------------------------------------- page 41 --------------------------------------------------

so often and demands so little effort from the student that the feeling of triumph quickly wears off.

5. The units give the student a sense of progress. The progress from unit 50 to unit 51 is a definite step forward, whereas a linear programme has so many frames that it makes little difference to a student if he has reached frame 560 or 580.

Thence we come

When we look today at language learning without a teacher, it may be interesting to recall what James Russell Lowell said in 1889 to the Modern Language Association of America: 'For nearly two hundred years no modern language was continuously and systematically taught here. In the latter half of the last century a stray Frenchman was caught now and then, and kept as long as he could endure the baiting of his pupils. After failing as a teacher of his mother-tongue, he commonly turned dancing-master, a calling which public opinion seems to have put on the same intellectual level with the other.' (Quoted from: J. R. Lowell, Latest literary essays and addresses. Vol.2, pp. 131f., London, 1892).

------------------------------------------------- page 42 --------------------------------------------------

Thither we go

There is no danger today that language teachers may become redundant because of the advent of programmed learning. Teachers will be needed by eager advanced students, and the programmes will ensure that more students than previously reach the advanced stage which makes them worthy of receiving the time of a live teacher.

Notes

1. The tremendous spread of language laboratories in this country can be seen from a report by the National Committee for Audio-Visual Aids in Education (NCAVAE) (in: VISUAL EDUCATION, July 1965, pp.4-15). The report states that there were then 452 language laboratories installed in this country. (It is not clear which installations are included in this figure. A breakdown of the total on p.5 says that it includes language laboratories installed for British Forces abroad, whereas on p.4, in addition to the 452 installations in this country, 22 are claimed to have been 'exported to Europe and developing countries (including British Forces abroad'.) Turner's list of language laboratories in Great Britain, which aims at being complete and up-

------------------------------------------------- page 43 --------------------------------------------------

to-date, only gives a total of 266 installations.

2. The NCAVAE report, p.9, contains tables implying that some secondary schools use the language laboratory for less than one period a week and some for only 1/4 of the total teaching time. No secondary school uses the language laboratory for more than 2 periods a week or for more than half the total teaching time. (In addition to these times, some schools make provision for individual study.)

J. C. Sager writes: 'Very few schools with laboratories can allow more than 10-15 minutes per language student a week, even if they use the equipment continuously throughout every hour of the day.' This sounds a very unlikely calculation. If a school has 15 language laboratory positions and each student spends 15 minutes per week at a tape-recorder, and if the language laboratory is continuously used from 9 to 4 for 5 days a week - as J. C. Sager suggests -, 2100 students of modern languages can use the language laboratory every week. Presumably a secondary school will also have students not doing modern languages. Surely there are few schools in this country all that large.

------------------------------------------------- page 44 --------------------------------------------------

It seems to be implicitly assumed by many educators in this country that language laboratory work must be beneficial, no matter how short or infrequent or irregular the opportunities are. However, it is not certain at all how much of students' reported and publicised achievements in oral fluency is due to the use, and to specific facilities, of the language laboratory. 'Satisfactory results' (normally no attempt at definition and measurement is made) can also be achieved without, and in spite of, the language laboratory. I believe that rather than use the language laboratory for a too restricted time or at irregular intervals, students should not use the language laboratory at all. (The usefulness of language laboratory busses wants reconsideration!) This is confirmed by an American experiment in which the results of three experimental groups working in the language laboratory were compared with those of control groups not using the language laboratory. 'Half of the experimental groups had practice sessions in the laboratory every day; the other half, only once a week. ... The results indicate that the students who had only one laboratory practice period a week *** made no more gains than those in the control groups; in one school the control

------------------------------------------------- page 45 --------------------------------------------------

group even made greater gains than the once-a-week group***.' (Hutchinson, p.14). The. problem is largely one of financing additional language laboratory places to give all students sufficient practice time. I have made some suggestions towards a solution in my paper 'Functions of a language learning centre and suggestions for its design'.

3. Rand Morton was the first to show that this is not so.

4. There is a tendency among teachers and writers on the language laboratory in this country to call all language laboratory work 'programmed learning', all tape preparation 'programming' and the contents of any tape 'programme'. This is not justified and highly confusing. Only materials at least intended to be teacher-independent in the particular field they cover should be called programmes. This article uses the term in this narrow sense. The majority of language laboratory tapes at present produced are intended to be only teaching AIDS and careful classroom preparation is necessary before the students can tackle these exercises. In contrast to proper 'programmes', these should be called 'exercises' or 'practice materials'. Cf. the definition of 'programmed instruction' given

------------------------------------------------- page 46 --------------------------------------------------

in L. N. Landa's paper 'Algorithmen and programmierter Unterricht', pp 37 f, and also the controversy between Günther (1963 and 1964) and Scharf (1964).

5. In my articles 'Uses of the Audio-Adapter' and 'Language laboratory work for advanced students and the Audio-Adapter', I have discussed by what type of work programmed learning can be supplemented in the language laboratory and how language laboratory facilities can be exploited or adapted for the purpose.

6. Observing students work on this programme draft has revealed many of its weaknesses and has led to the formulation of new principles for the construction of language programmes, which I hope to put into practice in an entirely new programme. My conclusions have been discussed in a paper 'Some thoughts on programming languages for the tape-recorder'.

7. See B F Skinner, 1960, pp 141 ff

------------------------------------------------- page 47 --------------------------------------------------

Sample of RU scoring book

NB
a. It is not normally necessary to devote 14 runs to one unit. This score has been designed to contain all the features that might occur in some score or other and is herefore rather lengthy.

b. The critical items

in run 2 are items 2, 4, 5, 6, 8, 9, 10;
in run 4 items 5 and 6;
in run 7 item 2 and item 9 independently;
in run 10 item 5 on its own and items 5 and 6 in collocation.

c. When the teacher checks the scoring books, he will first scan the 'totals correct' line. The totals should gradually increase from zero (or very small numbers) to 10. They should not stay the same or even decrease. The former means 'no progress' (= waste of time), the latter 'forgetting of items known'. The 'totals correct' line gives the warning signals that are easiest to spot. The teacher reacts to the signal by investigating the score more thoroughly, checking the recording script to see in what items the mistake has been made (this may lead to revision of the programme) and possibly by waking the student up or giving him a break.

------------------------------------------------- page 48 --------------------------------------------------

Glossary

Blank

Blanks are left to indicate wrong, or no, responses in a through run. (Wrong responses in intensive runs are not scored.) Blanks are used rather than crosses and the like for this purpose because their patterns are easier to see in the scoring book in contrast to the ticks and dots, and because they do not stress the element of failure in a wrong response.

Convenient target

Consecutive unmastered items in the same run are comparatively easy to locate on tape. Just like difficult items, they call for an intensive run once they have been spotted. (Diagnosis: Two or more consecutive blanks vertically).

Critical item

An unmastered item for whose sake an intensive run is done.

Difficult item

An item in which the student does not produce a correct response in spite of repeated attempts. To eliminate these as quickly as possible, 'repeated' attempts have been restricted to two. The difficult item is diagnosed by 'two consecutive blanks horizontally'.

------------------------------------------------- page 49 --------------------------------------------------

Dot

Used to score correct responses in an intensive run. Cf. tick and blank.

Drill

A violent way of referring to what would better be called 'exercise'. The popularity of the expression is in strange contrast with the audio visual movement, the hostility against 'old fashioned grammar' and the popular attempts at finding a more humane. and natural approach to learning and teaching. RU-programmes are decidedly ANTI-DRILL. (Cf. my article on 'Subvalidation' on the topic of natural stimulus/ response relationships.)

Group

(In an intensive run) Three or four consecutive items to be responded to in turns. Each group has at least one critical item , with non-critical items added to bring the group to the minimum size of three items.

Intensive run

Concentrated work directed at a group of items rather than a whole unit. It is followed by a through run as soon as the passing score has been reached in all groups (of a unit) necessitating an intensive run.

------------------------------------------------- page 50 --------------------------------------------------

Item

Corresponds to the term 'frame' in linear programming. Each item consists of a stimulus and a response and may also contain a 'Context Prompt'. (More about 'Context prompts' in my article 'Problems of learning in the language laboratory'.)

Non-critical item

A mastered item practised together with critical items to make up a group.

Passing score

The defined standard of performance which allows the student to proceed from one task to the next.

a. The passing score that permits the student to proceed from the test run of one unit is: 10 items (out of 10) correct in one run.

b. The passing score that permits the student to proceed from one group to the next in an intensive run is: Three correct responses for each critical item in the group.

c. The passing score that permits the student to proceed from one unit to the next (when the test run was not 100% correct): 10 items out of 10 correct in two consecutive runs.

------------------------------------------------- page 51 --------------------------------------------------

Repetitive-Unit programme (RU Programme)

A programme pattern that provides an alternative to the established 'linear' and 'branching' patterns and is especially suited to the programming of linguistic skills.

Run

The work performed by the student in his trying to respond to items in one unit.

Scoring book

A record of all the student's responses. It helps the teacher to control the student's working method and progress and helps the student to concentrate, eliminate wasted effort and reach the passing score as economically as possible.

Subvalidation

A testing procedure, to precede validation, whose purpose it is to assess not the teaching efficiency but the veracity of teaching material, the correctness and naturalness of stimulus / response relationships, the comprehensibility and unambiguousness of the visual and verbal stimuli and context prompts. The procedure was first suggested by me in 'Some thoughts on programming languages for the tape--recorder' and is discussed in more detail in my 'Subvalidation of language programmes'.

------------------------------------------------- page 52 --------------------------------------------------

Test run

The very first run through a unit. It corresponds to the 'pre-test' in established programming techniques. The results of the test run establish if the student can learn anything new from the unit in question or if he should proceed to the next unit immediately. This feature makes RU-programmes adaptable and highly suitable for revision purposes.

Through run

The work performed by the student by trying to respond once to each item in the unit. The first through run is called test run. Opposite: Intensive run.

Tick

Used to score correct responses in a through run.

Unit

Ten items demonstrating and demanding the same task to be performed. The 'little step forward' is from unit to unit, not from 'frame' to 'frame' as in established linear programmes. The student has to reach the passing score in one unit before he tackles the next.

Note added in March 2011: The term 'unit' was popular at the time when this paper was first published. It has now been replaeplaced by 'exercise'.

Validation

A testing procedure, generally used in programmed instruction, to assess the teaching efficiency of a programme.

------------------------------------------------- page 53 --------------------------------------------------

Bibliography

1 Bung, Klaus: Some thoughts on programming languages for the tape-recorder. In: International conference 'Modern foreign language teaching', Preprints Part I, pp.425-439. Berlin, 1964. - Reprinted in 7 below.

2 Bung, Klaus: Subvalidierung von Sprachprogrammen (Subvalidation of language programmes). In: PROGRAMMIERTES LERNEN UND PROGRAMMIERTER UNTERRICHT, vol. 3, no.1, pp.11-16, Berlin, 1966. - English version in 8 below.

3 Bung, Klaus: Uses of the 'Audio-Adapter'. First published in 7 below.

4 Bung, Klaus: Language laboratory work for advanced students and the 'Audio-Adapter'. In: CONTACT 8 (Fédération Internationale des Professeurs de Langues Vivantes), Vienna, June 1966. - Reprinted in 8 below.

5 Bung, Klaus: Problems of task analysis for language programming. First published in 8 below.

6 Bung, Klaus: Aufgaben eines Sprachlernzentrums and Vorschläge zu seiner Gestaltung (Functions of a language learning centre and suggestions for its design). Paper read at the '4. Nürtinger Symposion über Lehrmaschinen in Düsseldorf', March 1966. In:

------------------------------------------------- page 54 --------------------------------------------------

Lehrmaschinen in kybernetischer and pädagogischer Sicht 4 (Cybernetic and pedagogic approaches to teaching machines 4), edited by Helmar Frank. Ernst Klett Verlag, Stuttgart, 1966. - Translation in 8 below.

7 Bung, Klaus (Editor): Programmed learning and the language laboratory, vol.1. Longmac, London. To be published in 1967.

8 Bung, Klaus (Editor): Programmed learning and the language laboratory, vol.2. Longmac, London, 1967. (= The present volume)

9 Günther, K.: Entwicklung und Verwendung neuer Arbeitsmittel - ein Beitrag zur Erhöhung des Bildungs- and Erziehungsniveaus im Russischunterricht (Development and use of new teaching media - a contribution towards raising cultural and educational standards in the teaching of Russian). In: FREMDSPRACHENUNTERRICHT, vol.7, nos.7/8, pp.389ff., Berlin, 1963.

10 Günther, K.: Methodisch aufbereitete Tonbänder - ein Schritt auf dem Wege zur Teilprogrammierung des Fremdsprachenunterrichts? (Methodically prepared tapes - a step on the way towards partial programming of foreign language teaching?) In: FREMDSPRACHENUNTERRICHT, vol.8, nos.7/8, pp.354-365, Berlin, 1964.

11 Hutchinson, Joseph C: The language laboratory. How effective is it? US Dept.

------------------------------------------------- page 55 --------------------------------------------------

of Health, Education and Welfare. 0E-27021. Washington, 1964.

12 Landa, L. N.: Algorithmen and programmierter Unterricht (Algorithms and programmed instruction). In: Programmierter Unterricht and Lerntheorie (Programmed instruction and learning theory), published by Urania - Gesellschaft zur Verbreitung wissenschaftlicher Kenntnisse, East Germany, 1966. English translation in 7 above.

13 Marty, F.: Programming a basic foreign language course; Prospects for self-instruction. Audio-Visual Publications, Roanoke, Virginia, 1962.

14 Morton, Rand: The language laboratory as a teaching machine. University of Michigan, Ann Arbor, 1959.

15 NCAVAE: The use of language laboratories in Great Britain. In: VISUAL EDUCATION, July 1965, pp.4-15, London.

16 Sager, J. C.: Material for the language laboratory. In: AUDIO-VISUAL LANGUAGE JOURNAL, vol.3, no.1, pp.5-8. London, 1965.

17 Scharf, A.: Zur Anwendung der kybernetischen Betrachtungsweise auf die Gestaltung von Unterrichtsmitteln (A cybernetic approach to the design of teaching media). In: FREMDSPRACHENUNTERRICHT, vol.8, no.6, pp.298-303, Berlin, 1964.

------------------------------------------------- page 56 --------------------------------------------------

18 Skinner, B. F.: Teaching machines. In: Teaching machines and programmed learning. Edited by A. A. Lumsdaine and R. Glaser, pp.137ff., Washington, 1960.

19 Turner, John D.: Language laboratories in Great Britain 1965. London, 1965.

20 Twaddell, W. Freeman: On defining the phoneme. In: Readings in linguistics I, pp 55-80. Edited by Martin Joos. University of Chicago Press, Chicago, 1966.

top of page

Klaus Bung: Language Learning with Programmes

History of this article

Contents

Blank page

Restrictive practices

Encouraging developments

Audio-visual aids

Programmed materials

The role of the teacher

Language Gymnasium

Types of language programme

RU-programmes

Intensive runs

'Groups' in an intensive run

Scoring machine

RU algorithms

RU main algorithm

Logical conditions

Operators

RU sub-algorithm 'Do a through run'

Logical conditions

Operators

RU sub-algorithm 'Do an intensive run of 3 (or 4) items

Logical conditions

Operators

RU sub-algorithm 'Do item j'

Logical conditions

Operators

Insert: March 2011-03-17

Informal description of LASPEX-Preferred

Column 1:

Column 3

Column 4

Column 5

Column 6

Column 7

Column 8

Column 9

Column 10

Column 11

Column 12

Column 13 and 14

Advantages of the RU system

Thence we come

Thither we go

Notes

Sample of RU scoring book

Glossary

Blank

Convenient target

Critical item

Difficult item

Dot

Drill

Group

Intensive run

Item

Non-critical item

Passing score

Repetitive-Unit programme (RU Programme)

Run

Scoring book

Subvalidation

Test run

Through run

Tick

Unit

Validation

Bibliography

Klaus Bung:
Language Learning with Programmes