The author gratefully acknowledges the help of Dr. Paul V. Lorton Jr., University of San Francisco; and Rosemary N. Killam, Teaching Assistant, Stanford University, in making and executing numerous programming suggestions.

Historically, individualized instruction has been acknowledged as the ideal medium for all aspects of music instruction. However, undergraduate curriculum structure has tended to curtail individualized teaching of basic skills. In many instances, remedial tutoring of students on an individual basis has been the only method of helping them develop necessary facility with basic skills, but this requires a large amount of instructor time.

A decade ago the use of computers as instructional devices was an idea under consideration by only a few. Today, computer-assisted instruction (CAI) has undergone rapid development which is due in good measure to its potential for answering one of the current pressing needs in education: the individualization in instruction.¹

As early as 1954, in a paper entitled "The Science of Learning and the Art of Teaching," B.F. Skinner described the laboratory techniques that reliably produce modifications in the behavior of experimental subjects. His findings that the reinforcement of correct behavior is a necessary concomitant of learning led to his study of the classroom situation and the role of the teacher as a source of that reinforcement. He arrived at the conclusion that it is usually physically impossible for a teacher to provide every student with individual reinforcement each time it is appropriate and necessary.

Since much learning takes place covertly, it is impossible to know when a correct response has been made, an appropriate conclusion reached, or a correct idea produced. A tutor, on the other hand, interacts directly with one student, provides the guidance necessary to allow the student to arrive at a correct response, and then gives reinforcement. A tutor is always in the position of being able to provide necessary reinforcement and to adapt to the student's progress by further elaboration of ideas, repetition, and changes in pace.²

Thus, programmed instruction achieved its initial momentum because it appeared to provide the answer to the individualization problem. It fell short of reaching the ultimate goal, though, as its main contribution to the problem was to permit each student to move at his own pace. The computer-assisted instructional systems, on the other hand, provide individualized progress and individualized tailoring of the instructional interactions.

One of the areas in which computers have made great inroads is instruction in "drill and practice." Although the teaching procedures called "drill" have unpleasant connotations, there is little doubt that the accumulation of facts and elementary behavioral capabilities are prerequisite to advances at conceptual and integrative levels. Drill has a fundamental function in the learning processes. Licklider, in reporting on computer-based drill instruction, considered the following facts:

1. In present educational systems a large fraction of the total time and effort is devoted to drill.

2. It is inefficient to have 28 pupils sit idly by while the 29th reports what he just understood.

3. Neither teachers nor pupils enjoy the present kinds of drill enough to oppose its automation.

4. In drill, as in few other phases of teaching or learning, we can hope to obtain the masses of statistically homogenous behavior required to reveal the diverse effects and interactions we must have in order to understand the educational process.³

These considerations led to the development of computer programs that automate a procedure of drill. The computer has been found to be an optimum instrument for performing repetitive, time-consuming tasks. The CAI system can collect data during instruction on the errors made, as well as response time. Because of its speed in handling such data, the computer can use information collected to tailor the curriculum to the students.

Another important feature of the drill and practice mode is that it "converses" with the student by replying to his attempts to answer questions. Uttal, reporting on the conversational interaction in CAI, writes:

Consider the differences in the behavior of a student in a lecture hall as contrasted with the same student engaged in a conversation. The difference in the pattern of responses may be traced to the nature of the feedback which exists in each of these situations. In servoanalytical terms, the first case would be described as an "open loop" system, whereas the second is a "closed loop" mechanism with extensive feedback. The tutorial relationship is becoming increasingly rare as classrooms are filled with "open loop" students.⁴

Although educators frequently voice concern that computers could replace teachers, Dr. Patrick Suppes, Director of the Institute for Mathematical Studies in the Social Sciences at Stanford, presents reassurance that drill-and-practice systems are meant to supplement the regular curriculum taught by the teacher, who retains the function of introducing concepts. The role of the computer is to provide regular review and practice in basic concepts and skills. A big computer with 200 terminals can handle as many as 6,000 students on a daily basis in this mode. In all likelihood it will be feasible to increase these numbers to 1,000 terminals and 30,000 students. Says Dr. Suppes:

I want to emphasize that I do not see the computer as an instructional device in competition with the teacher. The role of the computer is like the role of books: to amplify the skills and time of the teacher. The economics of education will not, in any immediate future, change so as to reduce the student-teacher ratio at any level of education. Skilled teachers, by using computers to individualize instruction in the standard parts of the curriculum will reserve their own efforts for intensive individual instruction.⁵

Faculty and students in American universities have had considerable success in making application of computer technology to instruction and research. Forty per cent of the nation's colleges and universities had computer facilities by 1969. Many of these have been and more are being used for CAI.

In light of the foregoing history of the development of CAI, and the development of widespread computer facilities, teaching of music dictational skills has been a curriculum area holding great potential for the use of CAI. The previous stumbling block to development has been the lack of computer-controlled sound. The system described here has available to it computer-controlled sound, i.e., an electronic organ which can be played by the computer, which has proved to have a number of instructional advantages.

The basic needs leading to the construction of this system for CAI in music were:

1. Need for Sound

In the field of music instruction many teaching machines and teaching systems are in use which are extensions of textbooks and paper and pencil tests. These are adequate for cognitive learnings in music but do not suffice for developing musical skills, such as ear-training and dictation. For these latter a sound source is needed. That is to say, the primary stimulus, the principal mode of learning, should be provided by sound—not by a textbook, workbook, or by the terminal.

2. Need for Real-time Interaction

The student needs immediate feedback. On-line computer interface provides the interaction necessary for the student. Such interaction provides immediate reinforcement, either positive or negative, as there is instantaneous evaluation of the student answer. This is presented to him in terms of a positive comment if correct, and error message if incorrect, with some hints as to the magnitude and direction of each error.

3. Need for Individualization

Students need individualized instruction. Individualized instruction is increased through the use of various options. A student needs to be able to exercise the following options: he may select which strand of the curriculum on which he wishes to work; within any strand of the curriculum he may select a particular topic; within topics he may also select specific problems to work on. In addition to these considerations are those more traditionally associated with CAI: immediate feedback and individual tailoring of the curriculum.

4. Need for Student Records

Computer usage makes possible the keeping of detailed and accurate records of student performance. Such data are used principally in two ways: first, the usage of systematic errors of a student performance to select and construct supplementary material until it is mastered; second, the use of student data in summary form, to be transmitted to the instructor of the course, as a check on student progress.

5. Need for Research

The available data on student performance, when combined with the great flexibility of curriculum construction, makes it possible to use the system as a tool for basic research in the areas of how students learn and acquire these skills, as well as how to construct teaching strategies and curricula for effective learning.

 

DESCRIPTION OF OPERATION

The Music (Ear training) program at Stanford University is currently running on the PDP-10 Timesharing System of the Institute for Mathematical Studies in the Social Sciences. This system is dedicated to research in Computer Assisted Instruction and has many features designed to facilitate this research. These features include an elaborate system for enrolling and reporting on students' progress in a variety of courses as well as methods for collecting detailed data on the interaction between the student and the curriculum.

For the music program, the computer program interacts with the student through a Teletype® Model 33 KSR terminal and a Thomas solid-state organ model 145. The organ is connected to the computer with a 110 BAUD line through an interface which translates 8-bit patterns into notes to be played. At present the interaction with the organ is one-way. That is, pulses are only sent to the organ, and none are received from the organ.

The interface provides access to 64 notes on the organ. These notes are located in two octaves on the upper keyboard and three octaves on the lower keyboard of the organ. The 8-bit pattern also allows four functions to be set with each note: set to clear note, clear note(s), set to play note, and play note(s).

 

OPERATION OF CURRICULUM DRIVER

The curriculum driver for the music program is designed to give maximum use of the available features of the organ to the curriculum author. In addition, emphasis has been placed on having the program "generate" as much curriculum as possible thereby greatly increasing the leverage of the author.

Manipulation of the Organ

The organ can be "played" from the terminal or from the curriculum. To play the organ from the terminal in a free mode, the user types "PLAY" at the choice point in the program and then enters a mode of operation where approximately the same kinds of commands that the curriculum author can use are available. These commands include:

KBD	allows the switching back and forth between the upper and lower keyboard as the base keyboard.
KEY	allows the setting of keys for subsequent notes, chords, etc.
OCT	allows setting the base octave.
BEAT	allows varying the absolute time value of quarter notes.

Curriculum Handling

The curriculum is read by the curriculum driver from disk files stored on a Calcomp 215 Drive. Each strand in the curriculum is subdivided into topics. Topics are further subdivided into problems. Each problem normally consists of a problem statement and a correct answer. Optionally, hints and wrong answers can be stored with the problem. There are several possible formats which problems may have, and each of these is defined by a command to the curriculum driver which defines how the problem is to be handled.

The formats include:

I.	{"topic number"."problem number" (K)
	[< e > "problem text"]
	[< c > "problem text"]
	[K "key"]
	[S "set option"]
	[M "music text"]
	[C "correct answer text" | "response after correct answer"]
	[H "hint text"]
	[W "wrong answer text" | "response after wrong answer"]}

Format I is the basic format for standard frame-like problems. None of the parts of the problem are absolutely required. The brackets ({ }) separate problems from one another; the square brackets ([ ]) separate subparts of problems from each other; and the vertical bar ( | ) separates various fields of each subpart. Two special features are illustrated above. The "K" in parentheses after the problem number is used to signify that Kodaly notation is being used. The program then knows to expect designations other than the standard single letter notation for various notes. The "< e >" and "< c >", when they are present, are used to indicate elementary school or college level text statements.

"Key" and "set option" represent any of the key changing or organ setting options. These can be used to vary the setting of the organ for each problem as desired. The "key" RAN instructs the curriculum driver to select a key at random.

Within the problem text can be embedded music to be played during the presentation of the problem. When this is present, the terminal types the text up to the embedded music, waits to play the music, then resumes typing. This is very useful for illustrating and emphasizing explanatory points.

The following three formats are extensions of this basic format and all the possibilities of the basic format are available, when appropriate, in the remaining formats.

II.	{"topic number"."problem number"
	["problem text"]
	[K | "key"]
	[S | "set option"]
	[M01 "music text" | "correct answer text"]
	[M02 "music text" | "correct answer text"]

		.
		.
		.

[M10 "music text" | "correct answer text"]}

Format II is used to combine a number of musical passages into a single problem. This format, in a sense, abbreviates and cuts to the minimum essentials the tasks usually done under Format I.

III.	{"topic number"."problem number"
	["problem text"]
	[K | "key"]
	[S | "set option"]
	[G10]
	[G01 "music text" | "correct answer text"]
	[G02 "music text" | "correct answer text"]

		.
		.
		.

[G10 "music text" | "correct answer text"]}

Format III is an extension of Format II. The main difference is that the multiple subparts are played in random order and a criterion of 80% correct is required before the student can advance to the next problem.

IV.	{"topic number"."problem number"
	["problem text"]
	[K | "key"]
	[S | "set option"]
	[G | 10 | "triad or interval specification" | "answer type"]}

Format IV was developed to give the maximum leverage to the curriculum author. This format is most extensively used in the interval and triad strands (described below). The curriculum driver knows how to compute the notes in an interval or triad and generates several components of the problem at random within carefully prescribed limits. These generated components include the starting note of the sequence, whether the notes are to be played in ascending order, descending order or simultaneously. In addition, as in Format III, the order of presentation from the set is randomized and a per cent correct criterion is required before the student can advance to the next problem. As a final embellishment, the answer type determines whether the student is to type in the name of the interval or the name of the last note as the correct answer.

Response and Data Handling

After the program has presented a problem, playing whatever music is required, the student's response is awaited and evaluated. The student has the option (by typing a "Z") of replaying the music or of transferring to another strand, topic, or problem. If the student's response is correct, he is advanced to the next problem which can, under Formats II, III, or IV, be a subpart of the current problem. If the student was incorrect, the music is replayed and his next response is awaited. In addition, some attempt is made to indicate where his answer was incorrect. A sample lesson is included in the Appendix.

For each problem, the music played and the student's response is saved as data for later analysis. With this information the times of problem presentation and of the student's response are also saved. The detail of the data saved insures great flexibility in the analysis of students' progress through the curriculum.

 

DESCRIPTION OF CURRICULUM

The primary purpose of the system is to augment regular classroom teaching of music fundamentals at the college level. Curriculum entry was designed to correspond as closely as possible to music notation. Therefore, both pitch and rhythmic value of notes are entered in direct sequence, which makes for quicker debugging of curriculum entries and simulates as closely as possible the way in which a classically trained musician thinks.

Pitch names are entered by letter;

Sharp is indicated by +, flat by—(following the pitch designation);

Rhythmic values are entered by the first letter:

W = whole note, H = half note, etc. (with some minor variance in smaller values where 1st letter conflicts with other symbols). Numeral representation was not used because of the value conflicts in compound time signatures.

Default value is kept at quarter note—thus "C+" calls for C sharp, quarter note value. A rhythmic value without pitch designation indicates a rest.

The material to be presented is broken down into its smallest component parts, so that written instructions are kept to a minimum and the student is focused on the sound.

Maximum flexibility of curriculum choice is designed into the system, so that the student can, with one command, change to another curriculum area if he so desires, also so that he can skip ahead or review material already studied. However, if the student prefers to stay within a particular content area, the program advances him through an ordered sequence of graded material, evaluating his proficiency so that new material is not presented until a required level of accuracy has been reached.

 

CURRICULUM CONTENT

Separate areas of content within the program are designated "Strands." Content areas are, in general, those to be found in traditional ear-training programs. Thus, intervals, triads, rhythmic and melodic dictation, chord progressions, and modulations are treated separately within the program. The student is advised to switch from one related content area to another, as his skills advance. Within the content areas, curriculum is structured to be of advancing difficulty. If the student wishes, he may repeat any section, or he may skip a section or change to another strand, without being penalized. At any time, he may request that the material of the problem on which he is working be replayed, and he may request an infinite number of repetitions.

In addition, the student may speed up or slow down the tempo of the musical example, or he may change the key or the octave in which the example is played.

The following material is presently available:

Preliminary Strand—short optional beginning sequence

An introduction to music CAI; explanation of format and symbols used by the program. Two kinds of texts are used: "e" = elementary; "c" = college.

Interval Strand—for development of student interval identification and spelling

topics 1-7

identification of intervals according to MI2, MI3, PE4, DI5, MI6, MA6, MI7, MA7, PE8

topics 8-14

spelling notes constituting intervals, giving correct pitch of second note of interval with first note given

Melody Strand—graded sequence of melodic dictation

topics 1-5

movable "do"; presented in order and sequence of Kodaly System (elementary education level)

topics 6-8

melodic dictation of short, simple patterns

topic 9

melodic dictation with superimposed rhythms

topic 10

melodies with complex intervals

topic 11

melodies in minor keys

Triad Strand—used as review or introduction to interval drill

topic 1

root position: major, minor, diminished, augmented

topic 2

inversion: identification of specific triad quality

topic 3

review of first and second inversion, varying triad quality

topic 4

identification: any triad, any inversion

Rhythm Strand—graded sequence of rhythmic dictation

topics 1-7

dictation of rhythmic patterns and fragments, progressing from simple to complex patterns; pitch either held constant or varied

Chord Strand—graded sequence of dictation of chord progressions, nonmodulatory

topics 1-3

identification of tonic and dominant triads, as well as cadence patterns using tonic and dominant

topics 4-6

identification of all triads (by Roman numeral) in short chordal progressions of graduated difficulty

Modulation Strand—identification of key area to which chord progression modulates

short common-chord modulatory passages; student identifies by Roman numeral the new key

Experimental Strand—more advanced sequence, combining skills of Interval, Chord, Triad strands

topics 1-4

triads: identification of triad quality (major, minor, diminished) with required identification of inversion and Roman numeral in relation to tonic

topics 5-10

identification of 7th chords, inversions and Roman numeral in relation to tonic

topics 11-20

seventh chords, spelling chords with lowest note given

CURRICULUM DEVELOPMENT AND USAGE

The development of the curriculum listed on the previous page began with the dictation in Kodaly patterns, using sol-fa notation. The material was developed for primary school age children, and was tested with a group of students, grades 2-6. The material was found to be effective with this age level.

The second area of development was that of interval recognition. This strand continues to be the core of the curriculum, although students unfamiliar with CAI are now started with the Preliminary Strand.

Intervals are presented in groups of three contiguous intervals per topic. Pitch, direction, and harmonic/melodic presentation are randomly accessed by the program. The student is checked after his first five problems; if all answers are correct, he is directed to the next topic of intervals (which contains a one-interval overlap with the previous topic). If all of the first five problems are not correct, he is given another five, and evaluated to see whether he has obtained accuracy of > or = 80%. If not, he is presented with another series of ten problems, evaluated, etc., until desired accuracy is obtained. Student may ask for repetition of interval as many times as he wishes, or may choose to repeat a topic when it is completed. Or he may sign off in mid-topic (in which case he will be returned to the area from which he signed off at his next sign-on). Also, he may choose to go to material in an allied strand at any time. If he transfers to another strand, or back to a strand in which he was previously working, he is returned to the topic which he last completed, unless he indicates otherwise.

The subsequent development of curriculum was undertaken to augment areas of ear-training normally covered in a music fundamentals course. Some material was expanded at student request, and some specific and more advanced material was designed in cooperation with teachers whose students were using the program.

Since the program was specifically designed to augment the music fundamentals program, the great majority of students using the system are enrolled in the fundamentals courses.

In addition, time is made available to more advanced students who wish to review for departmental tests and to transfer students who wish to check out the program and make up any deficiencies in background.

During the '73-'74 fall quarter, sign-up sheets were posted weekly, and individual sign-ups limited to 20 minutes per session. (Data from the previous spring quarter indicated that to be an optimum length of time.) Once usage needs and patterns were established, students were allowed to sign up for regularly scheduled times on a reserved basis (reservation cancelled at first absence) with unreserved time allotted on first-come, first-served basis. In practice, heavy usage required that students plan to sign up at least 24 hours in advance for unreserved time.

Student response is remarkably enthusiastic. The system seems to be most heavily used by students who feel the need of more practice than can be presented in the classroom.

The flexibility and privacy of the system seems to be one of its major attractions. Students enjoy isolating their areas of weakness away from peer pressure; some engage in verbal monologues to the system dealing with their hopes and fears in the music program—would that it could be programmed for counselling!

 

SUMMARY AND PROSPECTS

The development of an ear-training drill and practice program is one of the logical outgrowths of CAI; historically, CAI has proved especially effective in drill and practice situations. Computer-controlled sound has proved to be the key to the program described in this paper. A number of alternate formats for material have been developed, which expand the resources of the curriculum writer. The program allows student control of a number of variables in the curriculum sequence. This aids in individualization of tailoring the material to the student, and provides immediate feedback and reinforcement.

The curriculum areas have been designed to correspond with the traditional outlines of music theory ear-training. The flexibility of the program has made it very useful as a supplement to the regular music fundamentals courses at Stanford. Student use has been heavy and response is very favorable.

The manner of entering answers on the terminals is influencing future development of curriculum. For example, entry of four-part harmonic dictation looks to be so cumbersome as to be of questionable value. The development of some off line material is being explored as a way to solve this problem.

Additional material which is presently under consideration includes modal recognition and dictation, 12-tone dictation, and identification of non-harmonic tones, all areas within the capability of the hardware of the system.

Appendix:
Sample Curriculum Output

HI

PLEASE TYPE YOUR NUMBER AND NAME
M1PAUL LORTON, JR.

JOB 4 ON TT1005 WED DEC 26 73 4:23PM-PST

HELLO, PAUL

HERE IS MUSIC SESSION 9
YOU ARE STILL WORKING ON THE INTERVALS & INVERSIONS STRAND TOPIC 1

INT 1 1

HERE'S AN INTERVAL STUDY SERIES.

TYPE THE RETURN KEY AFTER ALL ANSWERS.
FIND RETURN AND TYPE IT NOW.

.

NICE

INT 1 2

INTERVAL NAMES USE NUMBERS 1-7.

INTERVAL QUALITY ABBREVIATIONS:

MA=MAJOR
MI=MINOR
PE = PERFECT
AU = AUGMENTED
DI = DIMINISHED

TYPE MA NOW, AND LESSON PROCEEDS.

*MA

SMART

INT 1 3

TYPE SOLID THE QUALITY AND NUMBER OF INTERVALS.

EX: AFTER YOU HEAR THE FOLLOWING MAJOR 2ND, TYPE MA 2.

*MA2

GREAT

INT 1 4

NOW YOU WILL HEAR A GROUP OF MI2'S, MA2'S, AND MI3'S.
AFTER HEARING EACH INTERVAL, TYPE ITS NAME.
REMEMBER: TYPE Z IF YOU WANT TO REHEAR BEFORE ANSWERING.

*MI2

NICE GOING

*MA2

SMART

*MI3

SMART

*MI3

↑

TOUGH LUCK

LISTEN AND TRY AGAIN.

*MI2

GOOD, STUDY YOUR ERROR

*MA2

WELL DONE

*MA2

↑↑

MISSED IT

LISTEN AND TRY AGAIN.

*MI3

A BETTER ANSWER

*MI3

WONDERFUL

*MA2

GREAT

*MI2

SMART

*MI2

GROOVY

END OF INTERVALS & INVERSIONSTOPIC 1

 

---

¹See Joseph Margolin and Marion Misch, Computers in the Classroom (New York, 1968).

²William A. Deterline, An Introduction to Programmed Instruction (Englewood Cliffs, N.J., 1962), p. 11.

³J.C.R. Licklider, "Preliminary Experiments in Computer-Aided Teaching," in Programmed Learning and Computer-based Instruction (New York, 1961), p. 56.

⁴William Uttal, "On Conversational Interaction," in Programmed Learning and Computer-based Instruction (New York, 1961), p. 171.

⁵Patrick Suppes, "The Uses of Computers in Education," in Computers and Computation (San Francisco, 1971), p. 250.