Dear All, if you are running out of enthusiasm for density intensity, you might try this sentence from S. K. Heninger's sentenc from his HLQ article vol 50 (1987), p. 309: "The actual words Spenser uses are emnently forgettable."  He needs everey friende he can get.  TPRoche

----- Original Message -----
From: "James C. Nohrnberg" <[log in to unmask]>
Date: Friday, August 20, 2010 10:53 pm
Subject: Re: Spenser's lexical density
To: [log in to unmask]

> I appreciate the help here, re how to 
> calculate. It makes my own painfully 
> elementary position perhaps a bit 
> clearer: if an author used one word 10 
> different ways, semantically speaking, 
> and another author used 10 words for 
> the same thing, the second author 
> would come out as ten times as 
> vocabulary-rich as the first one (if I 
> understand what's been said).   The 
> first one would not get credit for 
> his/her kind of versatility (or 
> playing a sonata on only one string, 
> yes?).  ( --Some words are perhaps, in 
> themselves, vocabulary-rich ((or 
> lexically dense?)), as it were--see 
> OED for the verb "to take," on the one 
> hand, and German's lengthy compound 
> words, on the other.) Despite its 
> length, Finnegans Wake is the 
> vocabulary-richest text in the world, 
> is that right?  Gertrude Stein's 
> prose, roughly from the same period, 
> must be near the other pole (...is a 
> pole is a pole).  I think these 
> opposites in vocabulary storehouses 
> are related to contiguity vs. 
> similarity disorders (say, Stein's vs. 
> Whitman's or Hopkins' [my examples for 
> the term paper I wrote for I.A. 
> Richards, in 1961...agh.]).
> 
> Also: if a language is relatively poor 
> in rhymes, the long-distance rhymster 
> in that language might have to use the 
> same words several times over (at the 
> end of lines), thus diluting his 
> variety in words; or s/he might be 
> driven to resort to recherche 
> vocabulary to eke out a rime-scheme, 
> thus thickening it.  Would 
> particularly ingenious rhymes be 
> uncountable evidence for a 
> none-the-less rich vocabulary?  And: 
> an author gets no credit, in this kind 
> of counting, for words s/he invents, 
> or for abusio, or for remarkable 
> compounds, or for ten-dollar words, or 
> for ingenious alliterative 
> combinations of words.  Etc., etc. -- 
> Jim N.
> 
> On Fri, 20 Aug 2010 13:09:29 -0400
>  Kevin Farnham 
> <[log in to unmask]> wrote:
> > I happen to work with 
> >math/physics/statistics every day. 
> >Comparing equal numbers of words in a 
> >contiguous block is the only "simple" 
> >way to compare lexical density 
> >between two different texts -- i.e., 
> >the only way whereby you can have a 
> >degree of confidence that the 
> >resulting numbers are interpretable.
> > 
> > Also, the values for lexical density 
> >will always be a function of the 
> >number of words in the contiguous 
> >block: the higher the number of 
> >words, the lower the resulting 
> >lexical density.
> > 
> >For example, if we say:
> > 
> >   Lexical Density = [unique words] / 
> >[total words]
> > 
> > then the lexical density of "Ye 
> >learned Sisters" is 1.0. That, of 
> >course, doesn't tell us much, but it 
> >does illustrate that the smaller the 
> >size of the word sample, the greater 
> >the resulting density value.
> > 
> > Another example: say there were two 
> >authors who used the same 10,000 word 
> >vocabulary in their works. Author A 
> >wrote 100,000 words; Author B wrote 
> >500,000 words. If you divide their 
> >unique words by the total number of 
> >words in their oeuvre, you'd get a 
> >lexical density of 10% for Author A 
> >and 2% for Author B. Yet, really both 
> >authors have the same vocabulary.
> > 
> > So, the word counts have to be the 
> >same.
> > 
> > Even so, this simple calculation 
> >doesn't really tell us what I think 
> >people are really interested in -- 
> >or, what I think we're trying to 
> >measure with lexical density -- that 
> >is, the "word richness" or 
> >"vocabulary richness" that exists for 
> >an author's works. To get a better 
> >feeling for this, we might make a 
> >graph of the frequency distribution 
> >of the unique words, with the X axis 
> >being the word count, and the Y axis 
> >being the number of times each word 
> >appears in the text block. The words 
> >would be ordered on the X axis in 
> >order of frequency, starting with the 
> >most used word. Author A and Author B 
> >above might have very different 
> >graphs, though they employed the same 
> >10,000 word vocabulary.
> > 
> > By looking at these graphs, we'd be 
> >able to assess how "rich" the 
> >author's use of their lexicon is, 
> >albeit from a limited point of view.
> > 
> > Of course, we can also make that 
> >assessment by reading the works! But 
> >we find that dissatisfying, it 
> >doesn't give us a numerical index, 
> >and really we're all scientists in 
> >the 21st Century, so we like to 
> >measure art numerically (at least for 
> >fun)...
> > 
> > -------------------------------------------------
> > 
> > Craig A. Berry wrote:
> >> On Aug 20, 2010, at 12:52 AM, J. B. 
> >>Lethbridge wrote:
> > 
> >> 
> >> Which makes me less comfortable with 
> >>this whole approach to measuring 
> >>density the more I think about it. 
> >> If I were any good at statistics I 
> >>might know how to normalize these 
> >>numbers so texts of different length 
> >>could be compared, but as it is I 
> >>think comparing texts (or chunks 
> >>thereof) having equal length is the 
> >>only way to go.
> >> 
> 
> [log in to unmask]
> James Nohrnberg
> Dept. of English, Bryan Hall 219
> Univ. of Virginia
> P.O Box 400121
> Charlottesville, VA 22904-4121
>