A joint review by diverse hands, each writing one section
Setup [Sian Wiltshire, Bryn Mawr College.]
This review and installation was conducted on an Apple Macintosh (with 8 mb RAM and a 68020 processor), an Apple High Resolution RGB monitor, and an Apple CD SC. The System used was 6.0.5.
The Setup of Perseus 1.1 is for the most part quite simple to follow since it is laid out in a logical order: Hardware and Software Requirements, Installation, Startup, and Troubleshooting. It would have been helpful if part of the installation procedure had included the set-up of Settings found on the Gateway. Here, the user must chose the Path and Notebook stacks as well as linking Perseus with the Home card — very important — and the Videodisc player (if it is being used). The instructions for these procedures do not appear until Chapter 9 (Utilities) of the Manual. The Requirements are made quite clear. There is a problem with the definition of the word “optional.” The Videodisc player and monitor are “optional” since they are not required in order to use Perseus 1.1. SMK GreekKeys, however, are not really “optional,” but in fact a necessity if the user wants to work with the primary texts found in Perseus 1.1. In fact, no other Greek fonts can be used with this part of Perseus. The Installation is quite easy to follow. The authors collected into one folder, titled conveniently “Install me,” all of the parts of Perseus needed to run the program from the user’s hard disc. All the user needs to do is copy this information from the CD onto their hard drive. There is, however, some confusion over how to coordinate Hypercard 2.0 with the Perseus program. Since our office uses Hypercard with a variety of other applications, it was not feasible to place the Hypercard application nor the Home card in the Perseus folder as the instructions suggest. This transfer made it difficult to use other Hypercard stacks and applications. The simple solution is to make a copy of the Home stack (titled “copy of Home”) and place this in the Perseus folder. It does not seem to be necessary to copy the program of Hypercard 2.0 into the Perseus folder as the instructions for Perseus 1.1 indicate.
View 1: Click on Screen [Lee T. Pearcy, The Episcopal Academy.]
In a library one afternoon, an idle browser sits down at a vacant carrel and attempts to deduce its owner’s interests from the books left on its shelf. Classics, that is clear … but then? The browser sees a very thin outline history of archaic and classical Athens next to Loeb texts of Homer, Hesiod, Pindar, the tragedians, Herodotus, and other classical authors, as well as of Apollodorus, some of Plutarch, and Pausanias. There is an atlas of the Greek world and a Greek lexicon and grammar.
This carrel, then, belongs to a Hellenist. Its owner appears to be reasonably well-up in the language but not very well versed at all in the history of the sixth and fifth centuries, since the outline of history on the carrel assumes no previous knowledge of Greek history and is written in the clear, declarative English one associates with secondary school textbooks.
Then the browser notices a larger book called simply Art and Archaeology, and next to it a very odd book bearing the simple title Encyclopedia. Art and Archaeology contains photographs, drawings, and site plans with concise descriptions. These seem aimed at a different kind of user altogether from the one who was not entirely sure what Marathon was, what happened there, and why it mattered. The user of Art and Archaeology is at home with stylobates, raking cornices, Little Master cups, and polygonal masonry, or at least wants to know about these things. The carrel belongs, then, to a undergraduate student of archaeology, but one oddly uninformed about the most elementary aspects of Greek history and society.
The browser teaches students curious about ancient medicine and wonders what they will find in this Encyclopedia. No entry appears under “Medicine.” Looking up “Hippocrates” gives only bare references to Peisistratus’ father and three other characters of that name in Herodotus. “Physicians” draws a blank, as does a desperate stab at “Doctors.” “Drugs” produces only a reference to Medea’s magical herbs — but “Droop cup” has a 150-word entry with 15 cross-references. This sampling is not untypical. The Encyclopedia epitomizes the whole collection’s disparity between exiguous or elementary coverage of historical and cultural topics and detailed, specialized coverage of archaeological and artistic matters.
This is, of course, a virtual carrel, not a real one. It represents the Perseus 1.1 data and software, which other reviewers in this issue of BMCR will describe. Much time and labor has gone into the creation of this software. For the most part, it does what it is supposed to do. A user can summon up a map of Greece, locate Epidauros, see a plan of the Sanctuary of Asclepius, read a short encyclopedia entry about Asclepius, call up a Greek text of Pythian 3 and a translation, return to the site plan, focus on the Tholos or the Theater, locate information about them, see photographs of their ruins, and so on. The programmers of Perseus have done their job well.
Perseus also has designers. Whether they have made the artifact most appropriate to their medium remains a question worth asking. My answer is that they have not, and that their failure to do so can be attributed to the impeding grip of a very old image.
From Dürer’s St. Jerome in his Cell through recent times, nearly every picture of a student of the classics at work will contain one constant element: an arrangement of books. A text is on the table, and beside or below it a commentary. The student’s eyes focus on a line of text, and his finger touches the lemma tying commentary to text. The lexicon is at hand, and a classical encyclopedia. If the student is thinking about the Greek theater or the battle of Salamis, there will certainly be a map, and there may well be pictures.
Instead of this plurality of books, Perseus 1.1 offers a single compact disk containing data that can be accessed as text, commentary, lexicon, encyclopedia, atlas, and illustrations. Hypercard software allows a user to construct links among them. Yet Perseus remains an imitation of the student’s desk, with its little library of text and commentary, atlas and illustrated encyclopedia. With my description of the virtual carrel, I hoped to convey my sense that the data in Perseus are so odd a mix of the elementary and the specialized, and so full of gaps that it is hard to see what kind of students could make full use of the program in its present state.
Additional data in future releases, and perhaps a more judicious selection of data, will solve this problem. That correction will be laborious but is not difficult to imagine. Another, more obdurate problem will remain. It is hard to see what revision could exorcise from Perseus the ghost of Duerer’s St. Jerome, bent over his lectern, with those separate volumes ready to hand. St. Jerome in his study, with the new codices that were his information technology, flipped pages about as rapidly as Erasmus did, or as we do. The invention of printing spread information more widely among users but did not much alter the speed with which an individual user manipulated the information once it was in hand. The technology used by Perseus has transformed the saint’s study in two ways: by vastly increasing the amount of information available to an individual user at any one time, and by similarly increasing the speed with which that user can move among pieces of information.
Perseus exemplifies what might be called the horseless carriage phenomenon; that is, the tendency to use a new technology only to improve an old way of doing things. Its developers have assumed that what students of classics do is defined by the tools they have used for the last millennium. Students of classics sit at their desks, turn the pages of books, and make connections between items on the pages. Improvement, in this model, can only consist of bigger desks, more books, and faster flipping.
A sufficiently new technology, however, can create new ways of doing things. Although real advances in computer-assisted instruction will depend, I am convinced, on advances yet to be made in artificial intelligence, there is no need to speculate about massive parallel processing or six-inch cubes encompassing the contents of the Library of Congress. Computer technology in 1992 allows machines to accept input and produce output in a variety of forms, create mathematical models and simulations that respond to variations in parameters, and plot complex graphics on a screen.
Perseus seems intended for undergraduate students of classics. Here are a half-dozen things that such students cannot do, no matter how big their desks or how fast they flip their pages.
- Highlight a Greek word or a line of verse and hear it in Greek.
- See nouns in one color, verbs in another.
- Create new lines of Homeric verse by combining formulae.
- Superimpose profiles drawn by the Berlin painter on profiles drawn by the Pan painter.
- Given a demographic model of Attica in 417 B.C., observe how different values for birthrate, mortality, and age of females at first marriage affect the population of males able to serve in the fleet.
- Given a climatological model for Attica during the Peloponnesian War, observe the effects of different levels of rainfall on agricultural productivity.It seems reasonable to suppose that these activities or others like them might help undergraduate students of Greek language, literature, art, and history learn their subjects better. The technology used by Perseus could make these activities possible. Perseus itself does not.Future applications of Perseus technology, in fact, may find it effective to concentrate on small sets of data and forge complex, largely prescriptive links within these sets. As it happens, in the same week that I tested Perseus for this review I also examined Microsoft’s Multimedia Beethoven (1991). Multimedia Beethoven guides a student through Beethoven’s Ninth Symphony and conveys some information about the composer’s life and the historical circumstances surrounding the work’s composition. It is in many technical ways a far less complicated program than Perseus, and it uses its data in a far more authoritarian mode of teaching. It deals with a single musical work, not a civilization, and it presumes to tell the student how the symphony works and what it means. Yet if I saw two students in front of computers, one with Multimedia Beethoven and the other with Perseus, and had to say which student was more likely to be learning something, I would have to point to the student using Multimedia Beethoven. I would welcome a program that did for Sophocles and Oedipus the King what Multimedia Beethoven does for Beethoven and the Ninth Symphony.Reviewers act unfairly when they take authors to task for not writing a book that the authors had no intention of writing, and something of that unfairness may seem to have entered this review. Let me emphasize, then, an essential point: Perseus is not a book, or a collection of books. Books are what they are. The lines of a text proceed in only one direction. It is unfair, then, to take a text to task for having followed the only path available. Books are, in a sense, entirely data. By learning to read, we learn the instructions for handling all such artifacts, but what we are really interested in, what constitutes for us the richness of a text, is the richness of its data.The distinction between data and instructions, however, lies at the heart of even the most complex computer programs. (The difficulty of breaking down that distinction accounts for some of the obstacles in the way of advances in artificial intelligence.) A program’s designers create both data and instructions; an author, on the other hand, does not create the instructions we use to read his text. In evaluating a computer program it is reasonable, indeed essential, to consider not only the accuracy, depth, and scope of the data, but also whether or not the instructions represent an appropriate use of those data in view of the task addressed by the program.
The task of Perseus is teaching. Arrangement of information forms part of teaching, but it is not the same as teaching. In order for teaching to happen, one human intelligence must encounter another. A student using Perseus is not likely to come away with a sense that someone has intended that he learn something. Perseus has a team of able programmers behind it and a formidable committee of consulting scholars. Even so, Perseus lacks a unified authorial vision, and it lacks authority.
There is nothing wrong with a teacher’s eschewing the claim to authority. Socrates, as Gregory Vlastos has recently reminded us, did just that. But if disavowal of authority is to lead a student to knowledge, it must be informed by an active intelligence and by the intent that the student be so led. From that informed disavowal arise the ironic stance of Socrates, the distinctively Western tradition of scepticism about educational and other authorities, and our sense that in education the confrontation between the student’s intelligence and the teacher’s is just as important as that between the teacher’s and the student’s.
Much has been said and written about the inherently unauthoritarian nature of computer-assisted instruction, and the designers of Perseus have evidently taken this conventional wisdom to heart. Theirs is a neutral, liberal program allowing unguided links between miscellaneous data. We do not, I think, have enough experience with computer-assisted instruction to say for certain that it is an unauthoritarian medium. It may well be the case that the ironic stance of intelligent teachers in the tradition of Western humane education constitutes too fragile an algorithm for computer-assisted instruction to imitate. If so, the only possible good computer-assisted instruction may be inherently authoritarian, and the decision to avoid prescriptive links between data may be a weakness in Perseus.
The designers of Perseus have erred in taking the student’s desk as the reality-based model for their electronic instructional space, and the books that they have placed on their virtual desk form a collection with no clear intent or design for education and no particular student in mind. No matter; there are many reasons to praise them and their product. Their pedagogy may be old-fashioned, but their technology looks to the future and contains the seeds of its own improvement. There can be few tasks in our profession more complex than designing, organizing, and implementing a large package of computer-assisted instruction. Like most such packages, and unlike a book which, once published, exists in a single fixed form, Perseus can be updated and improved by new releases. Even in its present form, Perseus 1.1 is full of promise, and it will find users. More important, it may inspire others to explore genuinely innovative pedagogical uses for its technology.
View 2: Gateway [Richard Hamilton, Bryn Mawr College.]
What has most impressed me about the Perseus system is the system itself. The images are brilliant; the power and intricacy of the network are astounding. Though accessible within minutes its resources, like those of the MacIntosh on which it runs, are fully understood only after weeks of use.
My interest, though, is in Perseus as a research and teaching tool, and here my limited experience was mixed. Perseus offers two features not otherwise available: a database of Greek texts in English; a morphological analysis program. For either of these to be research tools, we must imagine scholars not able to read Greek, a prospect that turns my heart to stone. Perseus offers two kinds of data, texts and images, but for my research I have quicker and fuller access to texts through Ibycus and to images through the Beazley Archive and LIMC. Admittedly Ibycus will soon be available only in the Perseus/Pandora format and the Archive requires the extra step of looking up a photograph, which will not necessarily show the view I want, though I doubt many will find the “fifty or more views” of a given pot often offered by Perseus more useful than two views of twenty-five comparable pots. In the long run I think scholars will better served by limited but complete databases like the Beazley Archive, than by the relatively limitless but inherently incomplete Perseus. Then there is the copyright problem requiring considerable time for negotiation and placing fierce restrictions on much of the data.
Perseus can tell me things (especially with its Atlas) and show me things (esp. coins, which are enhanced rather than compromised by the space restrictions — though they sometimes inexplicably come out as clusters of gray dots — but also sites) but this helps me more as a teacher than as a scholar. For the teacher Perseus offers a great deal more, especially if one does not have a good slide library or Ibycus.
Perseus has many of the texts I teach in Greek (Thucydides, Herodotus, Homer, Sophocles, Aeschylus, Pindar, but not Euripides, Demosthenes, Lysias or Plato), and that means I can tell my Greek students that if they have a problem they can go to it for help with odd forms, but I will have to warn them not to let the translation that is provided do their work for them. I will also have to tell them how to go from the analysis to the dictionary — highlight the lemma — since the manual does not. (I have not yet figured out how to return directly to the text after doing an analysis. Nor have I figured out how the accents work on our keyboard. Documentation could be improved here.) My guess is that these students will not use it much.
My “in translation” students, on the other hand, almost without exception found Perseus extremely valuable when they were told to compare it to the Classics study in our library. I am teaching 38 students, about half freshmen, in a course that concentrates on the impiety trials in last quarter of the 5th C., and, although this subject and span are not well served by Perseus in either texts or images, the students were extremely grateful for having a single, understandable, speedy and enjoyable introduction to the Classical world. “It can help a student to find the right books and the right passages, and it can save enormous amounts of time by tracing one name or event through a bunch of sources.” “Keyword search gave me all references to Ephesus, and I was instantly able to read the sources Smith’s Dictionary only mentions… From now on I intend to use Perseus first for any research … all the information in Perseus is far more up-to-date than many of the library’s sources.” “While looking for one thing in Perseus there would be other names that would be eyecatching and curiosity-provoking so we would get sidetracked (procrastination potential here) and end up learning more than we had originally intended.” The photographs are “indispensable and helped to spark a curiosity I would not otherwise have had were the item merely described in the dry tone that seems to be the norm in scholarly writing.” “It was quite amazing to get all the references (via the English Word Search) on the screen so rapidly, and then to be able to click on them with the mouse to see the actual text.” “Because this information [about the Eleusinian mysteries] was so specific it was very helpful and informative; in the library, we were absolutely overwhelmed.” “Although I could have found the same information [on Aspasia in the library] as I did using Perseus, I would have had to haul my way vaguely through the ten or eleven primary sources that Perseus used. Although I would still use the library for getting the final information for serious research, I would definitely go to Perseus to get started.” “Being able to see what you want on a map without the clutter of a million other things about which you do not care at the moment is nice.” The message/option “this may take a while; continue search?” was “very nice for people with limited search time.”
These students were not entirely uncritical, but most of the criticisms pointed to the system’s obvious incompleteness. “After wasting about twenty minutes searching through various subject indexes and windows [for something on food] I nearly picked up the computer and threw it out the window.” “The maps and artwork are nice, but it would be just as convenient to look something up in an atlas or an art book. The memory that is spent on storing those incredibly beautiful vases and detailed maps would be better utilized in storing more books and journals.” “The exclusion of secondary texts was one of our largest complaints.” “The word-search option does not explain the context of listed entries.” One must resort to spelling creativity to find certain entries in a word search.” “The information [on women] was practically useless in comparison to the library.” “Hand-eye coordination becomes a key factor and the translated texts, as always, leave one at the mercy of the translators, as comparative translations are not provided.” “The manual was rather hard to follow.” “There were only two senences on Alcibiades and we could not find Socrates.” “The encyclopedia was extremely unreliable and useless … a large shadow was cast on the encyclopedia’s validity when we realized that it did not include Socrates.” “Euripides’ primary works were not even mentioned.” “There is a typo in Pausanias 6.15.10.” “When we looked up “sacrifice” on the computer, it told us there was no information, but in the libary, I found books on the subject.” The columns were “poorly drawn.” “There is no follow-up search, for example, where I could have asked to find all the instances where a particular word was associated with another word.” “What I need is a network with all articles published in Classics journals, listed by subject, period of history, primary literature used etc.” “The black highlight almost always expands to include the two lines above … every so often you get into an apparently endless loop of two or possibly three screens.”
Still these criticisms were set in the general context of praise, well summarized as: “Perseus has pretty pictures and succinct summaries to ease a would-be paper into existence; the library has the meat and potatoes that every growing paper needs to become a big, strong strapping sort of paper.”
We will be setting Perseus up in the foyer of our library as soon as possible.
View 3: A Cadd’s View [Harrison Eiteljorg, II, Center for the Study of Architecture.]
The Perseus system has no beginning, no end, no middle. It has pieces with neither explicit nor implicit order. It has connecting links, and it has information. One can use it, but one never knows for certain what has been missed. Therefore, I approached this review process not only as part of a group but with the idea that I would try to use the system as I think an advanced undergraduate student or a graduate student working at a beginning level might. I gave myself two asignments. First, I was to learn about the development of the scene building. Second, I was to learn about the government of Athens at the end of the Peloponnesian War. I thought the first assignment would test the system’s archaeological information base and its sophistication with that base. The second was intended to test the history and literature base.
The proper place to begin the session concerning the scene building seemed to me to be with the architecture material. That was wrong. I should have begun with the encyclopedia section. The terms were explained there, as was the development of the parts of the scene building. However, the terms were treated individually, and there was no single, continuous exposition of the nature, parts, and development of the scene building as a whole. The relationship between the theater and tragedy was briefly described. The most distressing aspect of this was the lack of bibliography. For a serious student, after all, this is a starting point; to be most useful it should point the user to the next level of information. Only Bieber’s The History of the Greek and Roman Theater was mentioned.
The architecural material was less helpful. There were only six theaters in the index — the Amphiaraion and the theaters at Delos, the Piraeus, Assos, Athens (Dionysus), and Epidauros. For each there was a plan, but the plans varied in detail and offered too little to be of much help. Each plan was available at small size only, with no capacity to enlarge for detail. Nor were there elevations to help inform the reader. For most of the theaters there were other illustrations. However, they were almost all unusable on the machine I was using. The one good photograph seemed to me to have false colors, and it was very small.
For the individual theaters there were bibliographic sources, but some of the choices for inclusion or exclusion seemed strange, since the sources for drawings were not necessarily included in the bibliography. For instance, the information screen associated with the plan of the Theater of Dionysus had the following legend: “Sources used for drawing © Perseus Project 1989 drawn by M. W. Cutler based on W. Dörpfeld and W. Wilberg (artists) in W. Dörpfeld and E. Reish 1986, in M. Bieber 1961 64 fig. 250, and on Travlos 1968 (artist) in Travlos 541 fig. 678” That seems to mean that the drawing was taken from Bieber’s reproduction of earlier drawings and from Travlos. Why was the original source for Bieber’s drawing not used? If they did not rely upon Travlos alone, why was there no bibliographical reference to any other source?
There were no apparent links between the theaters in the architectural section and the encyclopedia.
The result of my excursion to the theater was not what I had hoped. I had come away with only some very general information, no specific theaters with the parts well enough described for me to use them as examples, and inadequate bibliographic help to do more than follow up the individual theaters listed. Even with individual theaters, the bibliography seemed worrisome for the reasons noted above.
My attempt to learn about Athenian government at the end of the Peloponnesian War lasted over two sessions. During the first session I read the historical outline and checked some of the cross-references. The outline was very sketchy, indicating the numbers of the persons involved in the governments following the Sicilian disaster. However, little additional detail was provided. References to Thudydides and Plutarch helped to fill in some of the missing pieces (though in one case the reference indicated a council of elders, but the summary a council of 10 -with no seeming awareness of the problem and no explanation).
There were no references to other sources for more information, nor were there references to commentaries or other aids for using the ancient sources more fully.
I returned to look into the encyclopedia for more information about Athenian government. There were no links built into the system to suggest a connection between the summary and the encyclopedia; so I had no guidance as to the arrangement of the information. In fact, there seems to be no encyclopedia section on government.
During that second session, though, I did find some additional information about Athenian government in the historical overview. That would have helped a naive student reading about the changes at the end of the century, but the fact that there was no single place to read about Athenian government and no guidance to the places which did contain information came as a surprise. (The appropriate section in the historical overview was labeled poorly as well, since that section was said to deal with the reforms in the judicial system.)
It is difficult, on the basis of so little experimentation, to reach any general conclusions about the Perseus system. The experiment with the scene building was distressing in that the sources used were surprising and the potential for follow-up was limited. The number of theaters and the number of helpful illustrations were also disappointing.
The historical summary was just that, a summary. But instead of guiding the user to more complete sources of information, the system stops.
If this were intended for high school students, the absence of detail and bibliographic information would be acceptable, but the vocabulary used would not. For college undergraduates, on the other hand, the vocabulary is still demanding in places, and there is much of value to be learned. However, the material is still seriously incomplete, and the capacity for further in-depth study seems to be missing. For graduate students there seems to be only enough information to provide a start.
Despite the criticisms, the Perseus material is fascinating in many ways. Using it to try to find specific pieces of information, as I was, led me to bits and pieces which were new and interesting. That may be its greatest strength — that it contains so much material in one source. With all that information available at one’s fingertips, it’s easy to chase the odd line of thought. On the other hand, the biggest disappointment was the system’s lack of better bibliographic assistance. The bibliographic entries are not annotated, and the links between information and sources are too few.
View 4: Are you sure you want to exit? [James J. O’Donnell, University of Pennsylvania.]
The arrival of Perseus clearly marks a new stage in the e-ification of classics. Every living classicist saw the TLG disk in action sometime in the mid-80s and knew at once viscerally that this was a tool worth having: the success of that project explains why in the late 80s, classicists were generally in the van of humanities scholars in owning and using computers.
But Perseus is not TLG. Its potential reach is far broader than the 20 year old project designed to put linear texts sans variants on-line. To comprise texts of diverse kinds, photos of artifacts, and a variety of reference tools besides is an ambition several orders of magnitude grander than that of TLG. Accordingly, in these early Model T days of computing, the result is far less satisfactory in its present form. As now constituted, it is a teaching tool, barely a research tool. It is hamstrung by difficulties of copyright-protected access to existing materials (as indeed, n.b., TLG is if not hamstrung at least constrained), and by the primitive hardware and software with which we are still working.
In other words, the project is ambitious and far-seeing, and that is its gravest defect. My judgment is that the pricing is fair, and that the best use of this will be in undergraduate courses in institutions a little less rich than Penn or Bryn Mawr, where the value of the contents will be enhanced by comparative scarcity. My own preference would be to use it on a local-area or campus network in undergraduate courses in translation, and then to experiment with it in more advanced undergraduate classes. In those settings, it seems to me that a teacher willing to put in a little extra work will get a lot for her money.
One point needs to be emphasized, a sign of things to come. This is a resource that can be explored in a variety of ways. There is no single linear path through the museum, but an infinite number of possibilities. Even at the creepy speeds of a CD-ROM, progress is rapid, and mis-steps rapidly rectified. This kind of information accessible in this way has its greatest potential in giving the student initiative and opportunity beyond what the teacher forecasts. What I think I teach best is not ‘content’, that is what happened and when and what the text meant, but rather how to ask good questions. In teaching the language, what I concentrate on is making sure that in the presence of a difficult text, the student asks the right questions in the right order to get to an accurate comprehension quickly and easily. In teaching literature or history, I similarly emphasize the questions over the answers. Perseus elegantly facilitates this kind of teaching: the capacity for recording ‘paths’, i.e., sequences of text and image called up in succession, means that the student can answer the dictum ‘show your work’ as easily as in a math course. The teacher can analyze how assignments were done, and the student can herself feel quickly how well or easily a given line of thought is proceeding. Getting maximum benefit out of that feature in the present form of Perseus will take a little thought and doing, but seems to me the worthiest feature of all.
In the end, what are we comparing this to? It’s not a library yet, and won’t be for some years; but it has advantages over any textbook, and those advantages will increase rapidly with later iterations of the project.