Well Tempered IP Policy

In music one finds the concept of “temperament” in tuning.  Jim Loy has a nice discussion of the physics.  The basic problem is that the steps represented by the ratios of various notes, such as fifths (3/2), don’t exactly match octaves.  Moving through 12 fifths from C to C gives a second C that vibrates (3/2)^12 or 129.75 times faster than the first C, instead of the desired 128 times (2)^7.   The problem means that making the fifths follow a rule means that octaves will be off, and there will be some note combinations that don’t sound good at all.

There are various ways to deal with the problem, but it is inherent–there is a conflict between multiple systems of rules, so that things don’t come out perfectly by allowing any one rule to control, and it is impossible to require all rules to be followed simultaneously.  Instead, one has to temper the tuning in some way to impose compromises that allow some scales to sound “pure” according to expected ratios between notes, while other scales are not usable.  Or, one can anchor the tuning on octave divisions and mess with the notes within a target interval, such as the major third or fifth or octave so that they are off a bit from the ratio rule, but still sound nice, or interesting, just not necessarily “pure”.   As one moves away from a privileged key, such as C, one moves toward intervals that aren’t as nicely proportioned, or put another way, are proportioned in ways that add interest, not quite what the pure ratios governing intervals would predict, but to our ears, not so off as to be dissonant.

I have been considering such “well temperament” in another context, that of dealing with medieval English manuscripts.  I have worked for a long time on a text called Piers Plowman, a popular 14th century English alliterative poem that exists in more manuscripts than any other works besides Chaucer’s Canterbury Tales–and just ahead of the ever popular Lydgate’s Dietary.  Piers Plowman exists in three basic versions, called A, B, and C.  The manuscripts representing these versions, however, are a crazy mix.  There are A texts that have C or B continuations, there are versions where text from B is mixed back into the parts of A and then everything extended out, and there is a version Z which Walter Skeat calls “ridiculously corrupted rubbish” that makes its own sense (or nonsense) out of the poem.   Part of the puzzle is making sense of this manuscript history.  Given the 55 or so manuscripts representing 3 main versions of the poem in 10 manuscript subspecies representing a time of composition from perhaps the early 1360s to about 1393, if you want to read Piers Plowman, what version do you want?

I tend to stick to the A text.  It consists of two sections, a part called the “Visio,” which consists of two related dreams of society in its search for truth, and a “Vita” in which the dreamer embarks on a search fraught with allegorical adventure in an effort to find “Do-wel.”  Here is a sample of text from one of the A text manuscripts, T 3.14, of one of the more important arguments of the poem:  “When all treasures are tried, truth is the best”:

Here is a transcription:

Whanne alle treso/ur/s arn tri/gh/ed treu/th/e is /th/e beste

Where /ur/ is that squiggle over the “s” and /gh/ is a yogh, a g/gh character that looks like a script z, and /th/ is a thorn, a runic derivative that represents “th” (and is the letter represented by “y” in “ye olde ice cream shoppe” and the like).  Notice a few other things going on.  There are two forms of “a”, one with a round top (as in Wh.a.nne) and one with a looped top (as in .a.lle). There are two forms of “s” as well, one “long” and one “round” (tre.s.our.s.). There are two forms of “r” as well, though only the “straight” variant is shown here. “W” turns out to have no real distinction between capital and minuscule forms–and of course, since this manuscript was produced by hand, there were no “cases” to draw letter forms from, as there were in moveable type printing.

To play with representations of the text, I built a new font by sampling the manuscript for letter forms.  It’s still pretty rough since I haven’t worked out the kerning for letters and need to adjust the stroke weights, but here’s that same line, built out of standardized letters:

Now the text is a bit clearer, with the abbreviation expanded and more separation between letters.  Again, these letters are derived from the manuscript itself, creating a normalized “face” for transcribing, or writing any text (since the font is a true type font).  So, if we wanted, we could represent the modernization in the hand of the manuscript:

Still some spacing issues, but the general idea of the text is evident–one can bring forward a 500+ year old scribal hand and allow it to once again represent the text, but now with modern spelling.

Why do this?  Why the bother?  Well, one reason is that in Piers Plowman, it turns out that the character “Will” and the faculty of mind “will” both figure, and the manuscript does not distinguish these two with a “capital” for one.  So to read with the ambiguity, interpretative demand, and perhaps even willful playfulness of the author (who a lot of folks have decided  is named William Langland), one might actually want a font that’s suited to the task, that doesn’t push one reading or another.  But there’s another reason, not so evident from the line above.  Take this line (Prologue, 14):

Here is a semblable transcription:

Note the dotted y in “trighely” to distinguish it from the thorn.  And a modernized transcription:

I sai<gh> a tour on a toft tri<gh>ly i-makid

Or even more modernized:

I saw a tower on a toft “trighly” made.

What do we do with these words “toft” and “trighly”?  It’s critical to the alliteration on “t” but the meaning of the word for “hill” isn’t in current modern English and “trighly” is even worse at being modern.  Francis Covella translates the line like this:

I saw a tower on a hill artfully made.

It’s just that the line itself is now not “artfully made”–it’s rather a dull line that messes up the alliteration, and substitutes “artfully” for “trighly,” which swerves the line from the idea of well “tried” (or “tried and true”) to “art.”  Bother, if it is poetry.  Okay if it is an exercise in making an easier text to read.  But then one is reading a reading of the poem, and that’s a bit different in terms of the activity of interpretation.  It’s like, um, thinking about technology transfer by looking at AUTM’s statistics rather than experiencing the blowback from a jilted inventor directly.

In other words, the line in the manuscript has a more difficult word–toft–but the word is necessary to the alliteration.  There’s a conflict of rules.  One rule says, keep the alliteration and make readers look up words, and for that don’t bother to respell anything–make them look up trighly and i-makid, too.  Another rule says, if you are going to modernize, then you can’t leave some words unmodernized.  That would be inconsistent, and perhaps unfair to the modernized words, to have an old strange word among them, like “toft.”

No doubt you see where this is headed.  There appears to be different kinds of “tempering” of the text, by which one can move things a bit off from the rule and get a text that does some of both, not entirely consistent with each, but overall allowing the text to do more things–or the reader in working with the text–than if either rule was applied perfectly.

There’s a third thing going on here, and that involves disfluency.  Here is a discussion of disfluency by Adam Alter on The Edge.  If reading material is put into a form that’s harder to process, it changes the result:

We’ve looked at morality. If you present the same information in a format that’s difficult to read, either because the font that you’ve chosen is complex, or because you put the font against a background that isn’t very highly contrasting, or contrasted against that font, you find that people think that that moral transgression they’re reading is worse. By struggling to read the transgression, they basically assume that the transgression is worse. Somehow that ends up affecting or coloring how they feel about the transgression. If it’s much easier to process, if they can read it more easily, they tend to feel not quite as negative about the transgression. So this fluency principle has a whole lot of direct influences.

So if one creates a text and makes it easy to read, then perhaps one is changing the way in which the text is read.  I can introduce disfluency into a text by choice of font, by using old spelling rather than modern, by using a 500 year old scribal font, or by leaving in the text words that didn’t jump fully into modern usage.  It may be that the difficulties that are introduced aren’t the same difficulties that a reader in 1450 might have experienced, but it’s pretty evident that reading a handwritten text with multiple letter forms, abbreviations, and the like was not “easy”–the text was more like a kind of script than it was an authoritative stamp of characters dictating the reading.  A reader would have vocalized the words to recognize them, rather than rely on the printed shape of the words.  The glyphs make sounds, the sounds are recognized as language, the language is performed, and is a matter of recognition, interpretation.

Thus, a disfluent text might actually be a more interesting text, one tuned away from the ratios that make the key of C behave according to a set of rules about intervals–a text, with, one might say, more interest.  A well tempered text.

Here, the problem word is “shep”–it could be “sheep” or it could be “shepherd” and depending on what you decide, the poem’s narrator is either in disguise as a follower or as a leader.  Among modern folks, Skeat and Covella say “shepherd” but my colleague Vaughan thinks “sheep”.

All this stuff about medieval texts, temperament, and disfluency might seem some long way from university IP policies, but I would like to make the case that these concepts are very much in play in IP policies as well.  Yes, it may well be that printing IP policies in a disfluent font would make readers think their moral transgression is worse than otherwise, and that might be a good thing to do.

It does seem to be a bit more immoral, doesn’t it?

Deeper than that, though, in the architecture of policy, when a policy must deal with a number of issues and circumstances, as an IP policy pertaining to research discovery and innovation necessarily must, then there is a problem if that policy is set up to work in only one key, such as “commercialization that seeks to make money for the institution with as little interference or negotiation or consideration of circumstances as possible.”  A university IP policy has to be set up to deal with a number of situations–or scales–and has to balance, or temper as it were, these situations.  The stark reality is, such policies cannot follow a single rule set and be generally useful.

Imagine a university as an instrument, or as an old text with important things yet to say.  To force its tuning to a fit a primary scale, or to make it sound easy and simple, undermines the value of all its potential and complexity and opportunity, reducing it to a narrow practice and declaring a lot of other stuff dissonant, unauthorized, exceptions to rules made to look like defects rather than failures of the temperament, the architecture of the policy.

A university deserves a well tempered IP policy if it is going to have such a policy at all.  That means a policy that is not tuned exactly according to a single set of rules, but one that is off those rules in multiple directions, in order to make the opportunities across a wide range of possibilities available within the context of the policy.  For that, unfortunately, a policy that is inconsistent with some rules is better than one that is fully consistent with one set of rules but at a loss to all the other possible ways to play things.  That is why equity-based architectures for IP policies are more responsive to opportunities than are ownership-based architectures.  Equity-based architectures allow individuals and groups to make choices about the deployment and development of their work without requiring permission from an institutional authority to do so.  And given that institutional authorities, no matter how happy and diligent, end up having to represent institutional policies and practices and worries and interests, it’s just not possible for an institution to do what individuals can do, in the moment, and with personal priorities worked through.  An ownership policy displaces the personal and interpersonal and requires the creation of a new social fabric, one in which the institution is the most important player, not the inventor, not the principal investigator, not the industry volunteer who may know everyone worth knowing.

Thus, in the matter of considering university IP policies, my suggestion is that we need a better temperament.  We need policies that do not follow any one set of rules, or expectations, to the exclusion of others.  We need, in a nutshell, a pattern of judgment with regard to performance that allows a university’s resources to be played in all sorts of different ways.  The effectiveness of a university’s research and instructional programs is less a function of the focus of an office on “commercialization” as it is of the freedom of its personnel to explore a wide range of possibilities.  The value of the institution, and its IP workers, is in how these workers recognize possibilities in play and add resources as they are able, not that they are charged in pushing all possibilities into a narrow zone required by the architecture of the policy.

If universities want to move into a future that was dreamed of in the 1960s and 70s, where the university with its independent curiosity and readiness to adopt new things is again a leader of change rather than an increasingly expensive entitlement of the establishment, then they–faculty and administrators together–will adopt IP policies that are well tempered, and for that, a return to the concept of “equity” is a great place to start.

Wh2nne alle tresours arn trized treu@e is @e beXte

This entry was posted in Literature, Policy. Bookmark the permalink.

Leave a Reply