Exploring Poetry

The purpose of our project was to design, test, and evaluate several distinct software

programs, each on an individual poem. Each program would allow students in introductory

literature and poetry courses to explore the poem in depth and develop an interpretation of it.

With the help of a group of faculty, with whom we consulted throughout the project, we chose

the poems to be used, and in the first year of the grant tested preliminary and revised versions of

the first program, on Robert Frost’s “Design,” and completed the preliminary version of a second

one, on W.H. Auden’s “Musee des Beaux Arts.” In the second year, we presented our work at

three conferences, tested and revised “Musee des Beaux Arts” extensively, continued planning

and began the programming for two more programs (one on Gerard Manley Hopkins’ “The

Windhover,” and one on elementary prosody, called “Poetry to the Ear”) and hosted a conference

on “Computers and the Humanities” at our college. Students at a community college and a senior

college used the programs during the two years of the grant in a total of eighteen poetry or

introductory literature courses, and several other classes served as control groups. Many

suggestions made by these students and teachers were implemented in later versions of the

programs, as well as those of an outside evaluator on educational software design.


The problem our project addressed was the difficulty that students in introductory

literature classes often have with close reading of poetry, and their reluctance to do so. Our goal

was to emphasize close reading and analysis. During the course of the grant, our understanding

of the problem itself did not fundamentally change. In fact, the ways the students used the

programs made us aware of reading difficulties we had not anticipated. The records of each use

of the programs indicate how students go about reading and analyzing; this taught us more about

their learning processes than we could have discovered through traditional classroom instruction

(see “Evaluation” below for a description of these records.) The greatest problem many of them

have, we found, was in putting together the pieces of information they had learned into a coherent

overall interpretation. According to the data we collected, the programs were for the most part

successful in focusing attention on the language of the poems and allowing students to do very

detailed close reading. The specific questions they were asked to answer encouraged this, and the

nature of the computer itself commands concentration. Certainly most students enjoyed using the

programs. They particularly liked the ease with which they could access information about

individual words and concepts. They could work at their own speed, and this varied greatly from

student to student.

Any administrative pitfalls involved in using our software would have to do with providing

the requisite technology to support it. Technology is developing and changing at such a rapid

pace that both the software designer and the campus must constantly be prepared to adjust. The

college must be responsible for keeping hardware up to date, an expensive proposition in days of

tightened budgets. College technicians must be on hand for the multitudinous things that can go

wrong with the equipment. Teachers (especially those who are unaccustomed to using computers

with their classes) need some instruction in using the software, though the programs are really

self-explanatory and “computer literacy” is not required.

Those who embark on software development themselves will learn, as we did, that

everything takes longer than expected; equipment breaks down, unforeseen bugs occur, viruses

appear. Constant adjustments have to be made as one perceives how the programs are actually

used. For example, we discovered that each poem requires an entirely different approach and that

we could not simply copy the procedures we had already worked out. New developments in

technology allow for more and more sophisticated design and more and more possibilities, but

with each new invention adjustments in the plans have to be made. These are exciting times for

software development, but the speed of change in the field requires that expectations must

undergo constant revision.

An exciting outcome of our project was the unanticipated development of a method of

programming complex relationships through the input of faculty participants. Originally, we

ourselves determined the possible interpretations of a poem and decided how the students’

understanding of individual parts of the poem might fit into those interpretations. We found that

the program based on our judgments was to some extent faulty, chiefly in the appropriateness of

some of the comments the program made on the students’ answers. When we met with our

faculty participants, we realized that they brought to the poem in question a wider range of

understanding than we had originally conceived. We assigned them the task of developing a list

of possible interpretations of the poem as a whole. They discussed the lists submitted by each of

them at the meetings we held regularly, and, under Dr. Camp’s guidance, they ultimately reached a

consensus. We then designed a grid which the group would use to indicate what readings of parts

of the poem were to be judged consistent or inconsistent with the various interpretations.

After we administered the preliminary version of the program to the classes using it, we

asked the group to categorize the answers that students gave to specific questions, each expressed

by the students in their own way. Once the categories were formulated, we distributed the grid to

the participating faculty. On it, each teacher was to indicate the relationship between a category

of answer and the interpretations agreed on earlier. Once the grids were completed, discussed by

the group at large, and merged. Dr. Nimchinsky was able to use the resulting master grid to create

the database of responses and to modify the programming design. This method made it possible

for the program to respond to the answers which the students give in their own words and to

judge the internal consistency of their replies.

What we have described may seem a well thought out and organized approach to creating

a program dealing with text, and that is indeed what it turned out to be. In its development,

however, it evolved in stages that we at first did not notice or appreciate. Eventually, it became

clear that we had hit upon an extremely effective methodology for dealing with textual analysis

and interpretation. Moreover, we learned from this process that there are no short-cuts to

developing this type of software, that neither the programmer’s or the scholar’s ivory tower

provides a good vantage point for dealing with the reactions of others— especially students. We

feel at this point that, in the future, we can 1) teach our approach to other teachers and

programmers, and 2) extend this methodology to other fields in which the interpretation of texts is

crucial. In the meantime, we are eager to apply our method to other poetic works.