Author: Joe Page 9 of 31

This is another episode in my exploration of what we can see when we look at The Lord of the Rings through the lens of the frequency of French-derived words. Earlier posts are here and here, and the Silmarillion here. As always, kudos to the programmers of the OED Text Visualizer for providing the microscope.

We’ve seen that regular characters and situations use Germanic words. Like nuclear radiation, a low background level of French is unavoidable (viz., the louver on Meduseld), but as pomposity increases and sanity diminishes, the level of French in the text rises to double-digit percentages. This raises a question: does the craziest character use the most French words? The answer appears to be yes.

Gollum/Smeagol is almost certainly the least sane character in LotR. His dialogue needs quite a bit of editing before it can go to a computerized text analyser. The number of “s”s in a word needs to be standardized, sometimes “gollum” is an onomatopoeic punctuation mark not a proper noun, and so on. Also, his speech tends to be broken up in the text. The following computations are done on the closest dictionary-entry to Gollum’s words, aloud or internal.

There are seven blocks of dialogue long enough to support statistics between “The Taming of Smeagol” and “Shelob’s Lair”. One is a debate between Gollum’s two personalities; I’ve split that into its component parts. Depending on which of his personalities is dominant, the frequency of French words varies widely.

Frequency of French-derived words in Gollum’s speech

Tolkien gives us a brief flash of close reading from Sam that we can use as a guide: “[Sam] noted that Gollum used I, and that seemed usually to be a sign, on its rare appearances, that some remnants of old truth and sincerity were for the moment on top.” On each bar I’ve superposed the number of first-person singular pronouns in the passage. It correlates well, with the exception of one outlier.

The ring-maddened Gollum, as he talks to himself before he meets Frodo and Sam, reaches a level of French I’ve seen nowhere else in the text. If we recall our earlier estimate that something like 7% French is as far as a character can go without risking his health, Gollum’s 15.3% score is alarming. (Since much of that is his repetition of precious, maybe the computer isn’t telling us anything new.)

When Frodo uses his will, and the Ring, to dominate Gollum, Gollum’s word choices turn relatively normal for a chapter or two. Whether terrified or helpful, Smeagol’s French-level is healthy. But then, as the Gollum side recovers from the blow and he plots his revenge, he quickly blows past Feanorian levels into his record-level madness.

There’s one exception to this general rule. When Smeagol/Gollum gives us a short lecture on the history of Harad and Gondor, he briefly turns as normal as anyone in the book. He doesn’t use I, but he doesn’t sound much like himself either. And that’s a good thing. Later Frodo would say in another context, “his cure is beyond us; but I would still spare him, in the hope that he may find it.” There was no cure for Saruman, but Gollum could have been saved by an adjunct-lecturer position at the community college.

Case Study, Re-visualized

By Joe

On December 22, 2020

In Digital Humanities

Sara Waldorf’s master’s thesis at Signum University was built on a painstaking log of every use of the dative case in 2,000 lines of Beowulf. It’s entitled a “Case Study”, which is a joke so bad that Idiosophers have to salute. She gave a talk about this research in a “Thesis Theater” webcast at https://www.youtube.com/watch?v=VRhQHjIHSRQ.

Her thesis (available through the Signum U Library) contains the line, “The spreadsheet of dative categories could provide raw data for loading into a more sophisticated database.” Let’s take her up on that and see what we can do. Note: this post will contain no pie charts. They’re just empty calories.

It’s infuriating for the student but interesting for the philologist that grammatical cases are decaying during the Early Middle Ages. The obscure cases like locative and ablative were all collapsing into the dative case. The core of Sara’s work is unpacking these, leading to a classification of every dative into one of the 22 ways Old English used that case. Here’s what she found in these 2000 lines. Locative is on a line by itself because it’s by far the dominant function. The vestigial instrumental case, explicitly taught in the intro class, is almost completely gone from this sample.

		Locative	142
Dative of Means	68	State	20	Purpose	8
Object	53	Ablative	18	Impersonal	6
Dative of Reference	48	Reflexive	17	Measure	5
Indirect Object	47	Adverbial	13	Agency	3
Temporal	45	Cause	11	Inflected Infinitive	2
Possessive	32	Dative with Adjective	9	Modal	2
Accompaniment	26	Manner	9	Instrumental	1

Before and after the scribal change (click to embiggen).

Figures 4 & 5 in the thesis use pie charts to look at use of the dative in the 50 lines before and after the change of scribe. Here’s how I see them. As the thesis notes, there’s a huge difference. I’ve added the overall frequency chart for comparison. This enables us to see that the extraordinary case is the first scribe, who uses a lot of locatives. The second is much closer to the average for the whole selection.

The thing I liked best about Sara’s work (apart from the staggering labor of counting nouns and pronouns already being done for me) is how she split the selection up into separate stories. Beowulf is episodic. The way the poet talks changes with the material he’s talking about. This shows up in the dative-case usage. The stories in this sample are: the fight with Grendel’s Mother, the feast afterward, Beowulf’s departure from Heorot, Beowulf’s recap and prophecy about Freawaru’s marriage, presentation of gifts to Hygelac, the fight with the dragon, and Beowulf’s life in review. As good followers of Michael Drout, we should begin with a dendrogram: if we look at the fraction of all the datives in each episode that are of each type and cluster the episodes accordingly, they look like this. (complete-link cluster in 22 dimensions, four of which are boring)

Relationships of episodes by dative use

Among the seven episodes, there are three closely-related pairs. The gift-giving passage stands alone. Unlike the others, it’s loaded with indirect objects (without whom gift-giving is pointless).

The next step is to look at the density of various kinds of usage. Locatives are heavily concentrated in the first half of Beowulf’s departure. The scribal change is marked with a little orange dash in the middle of the departure episode. This is what we saw in the bar graphs above. The change appears to be due to the needs of the narrative more than the style of the scribe. The second scribe quickly returns to a locative pattern familiar from earlier episodes. Besides, had we used 100-line chunks instead of 50, the differences would have been minimal.

The dendrogram tells us that the Feast and Prophecy episodes are similar, but it doesn’t say how. Here we can see that they both start with some locative scene-setting, after which the need for locatives drops off. Grendel’s mother and Beowulf’s life in review trend the opposite way: few locatives at the beginning, but a double hump of them at the end. The gift section looks nothing like the others.

Density of locative datives

Indirect Object density

The deepest troughs of locatives correspond with peaks of two other common usages. (Naturally — you can’t have more than one or two datives per line, so when one type increases, the others must decrease.)

As we mentioned, the Gift-giving episode has lots of indirect objects. They fill in the trough around line 2175. The Departure and Dragon episodes, which didn’t look alike in the Locative graph, seem much more similar here.

The other low-locative part of the poem is at the beginning of the disagreement with Grendel’s mother. Those lines are full of Datives of Means. (A dative of means is like herebyrne hondum gebróden,”byrnie braided by hands” in line 1443.)

Density of means

There is a moderate surge in datives of means toward the end. I was expecting that to be inflated by pairs of nouns: wigum ond waépnum in line 2395, for example. (He supported the son of Ohtere with warriors and weapons.) But Sara knew that was coming, and only counted such doublets once. These peaks indicate many successive sentences using datives of means.

The densities of the other 19 uses of the Old English dative are less obvious in their meaning, but are available for consultation here: DativeDensity

Conclusion

I shall not pretend that the motivation of this study is anything other than, “Someone made a database — let’s look and see what’s in it!” Sara’s conclusion relates to the transition during the Anglo-Saxon period from inflection to preposition as a way to indicate the dative functions in an English sentence. My interests lie rather in seeing the ebb and flow of grammatical structures in response to the narrative. I would have expected each of the episodes to begin with a burst of locatives. (“In a hole in the ground there lived a hobbit.”) It’s interesting that most of these don’t. Though the poem as a whole begins with a burst of datives, they’re temporal and accompaniment; none are locative.

The “agglomerated dative” offers one way to quantify the presence of grammatical structures in the poem. Whether it can be expanded to yield new insights into the perennial questions about Beowulf remains to be seen.

References

Drout, Michael. Tradition and influence in Anglo-Saxon literature: An evolutionary, cognitivist approach. Springer, 2013.

Waldorf, Sara J. “A ‘Case-Study’: Functions of the dative in Beowulf lines 1439-2439”. Signum University, 2019.

On the sentience of the Ring

By Joe

On December 17, 2020

In fantasy, squibs & crackers

Tom weighs in on the question, “is the One Ring sentient?” with some evidence that the answer is “no”. The nerds on Reddit had an interesting discussion about the post. (Sturgeon’s Law applies, of course.) The gang raised a good question about the wheel of fire talking to Gollum, for instance.

It’s tricky, because the word “sentient” isn’t often used according to its dictionary meaning. It means “sensing the world around it”, but people generally use it to mean “thinking”. Until we get to Mordor and the hallucinations start, only the former seems relevant.

A good lens through which to look at Tolkien is to look for real-world analogues of what we see the Ring doing. Fact: It changes size, to get away from its current bearer. Tolkien is careful to say “seems to” all the time, but the physical evidence seems clear. The Ring came off Isildur’s hand, and Gollum’s hand, and it tried to escape from Bilbo several times, and boy did it hate being anywhere near Bombadil! It grew as big as it could in a vain attempt to get out of his palm.

So in some way the Ring knows when it’s not going to get anywhere with its current bearer. It knows when an opportunity for something more congenial comes around. And it can change its shape accordingly. Is there a real-world analogue for this?

Seeds have one. They can sense moisture, temperature, and gravitational potential as gradients around them. When they get the combination of moisture decreasing, temperature increasing, and gravitational potential increasing all in the same direction, they sprout, and send a shoot that direction to get out of the dirt and into the sunshine. This is parallel to what the Ring is doing, if we can find some field around it that relates to Sauron’s power instead of earth and water. Some kind of luminiferous aether, except for the power of the Ainur. Ilmen, perhaps? Could Sauron or the Nazgul be distorting the density of ilmen as part of their attempts to draw the Ring to them? (Of course, the palantir can do something similar, so maybe it’s something more down-to-arda than that.)

Anyway, the gradient of the whatever-field affects the Ring’s size. When it’s near a person more congenial to Sauron’s goals, which could be someone more powerful and closer to evil, or someone less powerful but less good, it expands. This ties in with the complexity of Gollum’s character — he’s neither good nor evil, not really anything except lust for the Ring, so any random goblin would be a better host, and off the Ring fell.

P.S. Anyone who doesn’t like the idea of Sauron creating a thing with the power of a seed (as I’m sure JRRT wouldn’t) is invited to use a slime mold as the model instead.

Of Beleriand and its Vocabulary

By Joe

On November 11, 2020

In Digital Humanities, fantasy

Daniel Stride, antipodean writer, book-blogger, and remarkably-good US election forecaster, wonders if Denethor’s french-fried vocabulary applies to Fëanor’s mad rants, too.

There are some methodological problems here.

The Silmarillion isn’t a finished work.
The Silmarillion has multiple authors.
Fëanor doesn’t make any sane speeches for comparison.

But what the heck. Full speed ahead. The longest verbatim speech we get is in Chapter IX, “Of the Flight of the Noldor”. I won’t repeat the whole thing here, but it runs from “Why, O People of the Noldor…” and runs to “No other race shall oust us!” We can validate our presumption that this is a mad rant by noting that 11 of the last 13 sentences end with exclamation marks.

There are 27 French words out of 299, or 9%. The French words in Fëanor’s rant almost suffice to convey the entire meaning of the speech: race, oust, pursuit, beauty, levels, endure, ease, return, realm, people, vengeance, cowards, mountains, await, conquered, jealous, folly, city, journey, serve, war, regained, enemy, valiant.

This raises the question: How close is that to Denethor’s 10%? Most English prose will be loaded up with articles and prepositions that are all from Old English, so there’s a ceiling on how French a passage can be. Is 9% significantly different from 10%? What is the range of tolerable frenchification?

To look into that, I processed the first paragraphs from Chapter XIV, “Of Beleriand and its Realms”, which has to be the least-crazy part of the Quenta Silmarillion. That came out to 20 French words out of 266, or 7%. The words are ages, borders, fortress, defence, assault, dungeons, war, haste, destroy, search, tunnel, issued, mountains, furnaces, refuse, issued, desolation, plain, citadel. Morgoth definitely skews toward the Romance languages. Note the dungeon and the tunnel — they’ll be back shortly.

Who else isn’t crazy, besides the Narrator? Beren comes immediately to mind. He doesn’t make any long speeches [1] that I recall, but if I splice together everything he says to Thingol on their first meeting it adds up to 200 words. Of those, 10 are French. Just 5%: perils, possess, jewels, powers, spy, price, perform, fate, rock, crown. (Not such a good plot summary, but it catches a lot of the flavor.)

In case anyone is wondering if the human/elf distinction matters, I decided Legolas was a sane elf. His speeches in LotR range from 2% French when he’s singing a lament for Boromir to 11% when he’s talking about visiting the Glittering Caves. [2] His average is 6±2.5%. In general, the Quenta Silmarillion has a higher French quotient than LotR, but they’re comparable. Christopher Tolkien and Guy Gavriel Kay took care to match J.R.R. Tolkien’s style and they picked up on this facet very well, arithmetically speaking.

So, with Sam’s permission we will call that settled. As long as they can stay above ground, sane people in Tolkien use less than 7% French words.

[1] This may be a better sign of sanity than any amount of etymology.

[2] It’s almost impossible to use English words to talk about caves.

Descent into Madness

By Joe

On November 7, 2020

In Digital Humanities, fantasy, Vive la France

The team at the Oxford Dictionary have upgraded their text visualization tool. The first beta version was the tool of some idiosophizing a few months ago. This new version still has the 500-word limit, but it’s gotten better at guessing which meaning of a word the author had in mind, and it handles Elvish words and proper names much more gracefully. That is, it ignores them.

The madness of kings and the damage it can do to a country has been on my mind of late, so today I used the new tool to look at Denethor’s first and last speeches. Long-time readers know I’m an admirer.

Here’s the first thing of any length we hear Denethor say, after removing all the things that aren’t Denethor:

Dark indeed is the hour, and at such times you are wont to come, Mithrandir. But though all the signs forebode that the doom of Gondor is drawing nigh, less now to me is that darkness than my own darkness. It has been told to me that you bring with you one who saw my son die. Is this he? …
Verily. And in my turn I bore it, and so did each eldest son of our house, far back into the vanished years before the failing of the kings, since Vorondil father of Mardil hunted the wild kine of Araw in the far fields of Rhun. I heard it blowing dim upon the northern marches thirteen days ago, and the River brought it to me, broken: it will wind no more. What say you to that, Halfling?

As before, the size of the circle is how common the word is in English, the horizontal position is the year the word entered the language, the vertical position is how many times the word appears in the text, and the color of the circle is the language family whence the word came into English. Blue is Germanic (dark for Old English, lighter for German or Norse), red is French (and other Romance languages). Other languages appear in other colors, but these passages don’t have any of those.

Visualization of Denethor's first speech

This is how a great leader of men talks

And here is the last speech Denethor makes before he ignites the pyre, similarly edited:

Pride and despair! Didst thou think that the eyes of the White Tower were blind? Nay, I have seen more than thou knowest, Grey Fool. For thy hope is but ignorance. Go then and labour in healing! Go forth and fight! Vanity. For a little space you may triumph on the field, for a day. But against the Power that now arises there is no victory. To this City only the first finger of its hand has yet been stretched. All the East is moving. And even now the wind of thy hope cheats thee and wafts up Anduin a fleet with black sails. The West has failed. It is time for all to depart who would not be slaves.
Hope on, then! Do I not know thee, Mithrandir? Thy hope is to rule in my stead, to stand behind every throne, north, south, or west. I have read thy mind and its policies. Do I not know that you commanded this halfling here to keep silence? That you brought him hither to be a spy within my very chamber? And yet in our speech together I have learned the names and purpose of all thy companions. So! With the left hand thou wouldst use me for a little while as a shield against Mordor, and with the right bring up this Ranger of the North to supplant me. But I say to thee, Gandalf Mithrandir, I will not be thy tool! I am Steward of the House of Anarion. I will not step down to be the dotard chamberlain of an upstart. Even were his claim proved to me, still he comes but of the line of Isildur. I will not bow to such a one, last of a ragged house long bereft of lordship and dignity’
I would have things as they were in all the days of my life and in the days of my longfathers before me: to be the Lord of this City in peace, and leave my chair to a son after me, who would be his own master and no wizard’s pupil. But if doom denies this to me, then I will have naught, neither life diminished, nor love halved, nor honour abated.

Not the words of a well man.

When we meet him, Denethor is sane. He is firmly in command of the defense of the West. In a speech of 140 words he uses only three of French origin: sign, river, and march. Those are good, short words. You can barely fault him.

But the palantir is a dangerous thing. One must have a great strength of will to use it without being deceived. Mighty as Denethor was, madness took him, and we can see it in his speech: 37 French words out of 370. Alas for the son of Ecthelion! Sauron’s lies turned him 10% French, and from that there is no return.

Caveat

The results of the visualizer can be sensitive to the date of the text. As you can see from the vertical red lines, I chose a date before the author started writing, not the date of publication. Using the later date means there’s a risk that the robot will discover some obscure technical term that came into existence just before the book hit the shelves. The OED knows everything, and that has negatives as well as positives. (In this case, the tricky word was “kine”.)

I didn’t know where my towel was

By Joe

On October 4, 2020

In squibs & crackers

It often happens while I’m doing manual labor that I get to wondering about a word. Today the word was “hitch-hike”. I remembered a story about its origin, and decided to check it out during my next break. The story was not true. But then I found this citation in the Oxford English Dictionary:

We may charge this wicked hitch-hiker the ten cents extra that she deserves for asking for a bath towel.

It’s from Barbara Starke, Touch and Go: the story of a girl’s escape. 1931. I can’t find this book online — neither the Internet Archive nor Google Books has it. You can buy a used copy from Amazon for $847, if you like. The references I’ve been able to find make it sound interesting — it was banned in Ireland.

Some humorists think of jokes from reading the newspaper, some from advertisements, some from pop culture. I think I now know where Douglas Adams got his.

I feel much hoopier

How to make a dictionary

By Joe

On September 26, 2020

In Slow learner, Vive la France

In which we once again find our Idiosopher using insanely-powerful Internet research tools for frivolous ends.

A while back, Prof. Emily Steiner, who seems to be familiar with every medieval manuscript that’s survived to the modern era, tweeted an image of what she asserted to be a debate between a cockatrice and a wyvern. It’s a snippet from Brunetto Latini, Livre du Trésor (1230?-1294). Manuscript BnF Fr 568, folio 48v (available via Gallica).

I can’t resist reading the captions on medieval illuminations, just to see if I can. The red letters looked to me like “De toutes maines de serpens”. Obviously this means “Concerning all kinds of serpents”, except for the word “maines”– what’s that?

“Why did it have to be snakes?” – Prof. Henry Jones, Jr.

I consulted the Dictionary of medieval French, and found definition 2 for “maine” is “manière, espèce”, with a citation to Le Roman de Tristan. So, fine. [1] But there’s one thing I overlooked: that red curlicue over top of the word I’m puzzling over is a scribal abbreviation. After a couple of us got ourselves confused, Prof. Steiner let us know it stands for “re”, and the pen-strokes I read as “in” are actually “ni”. Properly read, that word is actually “manière” itself, and it hasn’t changed in 800 years.

That implies a nuance that hadn’t occurred to me about the lexicographer’s art: the scholar who wrote the dictionary included a word he knew didn’t exist and wasn’t used, just because it’s easy for students to read it that way in the text. Awfully considerate of him.

[1] The Romance of Tristan is attributed to an author called Béroul, about whom nothing is known. The manuscript is in poor condition. The dictionary tells me the word “maine” is used in the phrase “male maine”, meaning intransigence or evil will, not bad manners. I tried to find the word in the manuscript, but failed. Maybe it’s underneath one of the coffee spills.

A pretty high-level piece of Tolkien geekery

By Joe

On September 13, 2020

In mythgard academy

The title of this post refers to one of the Idiosopher’s suggestions for the school-spirit section of the bookstore. It is the opinion of President Corey Olsen of Signum University. Praise is always welcome, especially from one of the world’s premier Tolkien geeks.

Headley’s Beowulf

By Joe

On September 1, 2020

In Alliterative Verse, squibs & crackers

I have just finished reading Beowulf: A New Translation by Maria Dahvana Headley. Not without some trepidation, because the reviews in the mass media and on Twitter all advertised it as a translation into current Internet idioms. That could have been awful, but I’m here to tell you it was all just clickbait. This is a “new translation” in the sense that the English used here is from the last half century. Sure, the pull quotes that they quoted are in there, but they’re not important to the text. One of the most popular lines to quote looks frankly spliced in to attract attention.

This is what I was afraid of

The words in the book are much better than the ones in the reviews. MDH uses the difference in language between the current world and standard modern English as a tool. The narrator, as I hear him[1], is the same guy who narrated the original. He’s old and getting a bit cranky. He puts in some contemporary locutions to attract the kids’ attention, but the moment he’s got it, he slides back into an archaic mode. That contrasts with the more colloquial sound of the direct quotations, forming dissonances that are sometimes brilliant:

Only then did Hygelac begin to question his comrade, calmly, commandingly, to glean the story of the war-Geats, and take the tale for his own hall-history.

“Holy hell, Beowulf, how’d it go out there?” (1988)

One question I always have to get out of the way in an adaptation of Beowulf is, “Did the translator do her homework?” Not even a question here. In the introduction, she deftly gets Tolkien out of the way, in accordance with Terry Pratchett’s dictum.

So I definitely liked this book. It is the fastest to read of all the versions I’ve read. A backhanded shot at Princess Diana was completely unexpected, but worked well, too.

[1] I’m sure “him” is still the right pronoun. That’s not the case for all the characters in this book — MDH is good at making little gender-bends that stay within the original text.