Śabdānuśāsana – Explanation of words

[artkeidan]

Lees maar, er staat niet wat er staat

This is the most famous verse by the Dutch poet Martinus Nijhoff. Literally translated, it says:

Just read, it doesn’t say what it says.

I see this as a perfect metaphor for what a poet — and more broadly, any writer — truly does.

[artkeidan]

First (un)expected theoretical problems

I decided to create this online application, IEI, which displays the regular correspondences among the Indo-European languages and allows users to visualize isoglosses.

I did this partly for fun, but also because transforming an idea into a formal algorithm (and every application is, ultimately, an algorithm) is one of the best ways to test its internal consistency and uncover hidden contradictions. Above all, I wanted to see how well the Indo-European hypothesis would withstand formalization.

The first difficulties did not take long to emerge.

Consider Grimm’s Law, specifically the part according to which Proto-Indo-European voiced aspirated stops became plain voiced stops in Germanic. Have you ever wondered what this statement actually means from a phonological perspective?

In phonology, only distinctive features matter. The Proto-Indo-European phoneme *bʰ can be characterized as {stop, voiced, aspirated, labial}. The Gothic phoneme b, which derives from it, can be characterized as {stop, voiced, labial}. But is Gothic b also “unaspirated”? No, it is not. In the phonological system of Gothic, aspiration is not a distinctive feature at all. Therefore, describing b as non-aspirated is meaningless. It would be like saying that b is non-red or non-fast.

This is precisely how phonology works. If a feature is not employed by a given phonological system, it does not acquire a negative value; rather, it ceases to be relevant altogether.

It is far from obvious how to formalize the logic underlying rules such as those of Grimm's Law without resorting to some ad hoc workarounds.

I am nevertheless confident that a satisfactory solution can be found, and I continue to work on it.

[artkeidan]

Glottalic vs. Classic PIE: an asymmetry

I've been building the Indo-European Isoglosses explorer, an online database that — among other things — displays the phonological inventories of the various IE languages and traces the PIE origin of each individual phoneme.

The project has a section that displays the PIE phonological inventory itself. Simple enough, right? Except: which PIE inventory?

There are at least two serious options on the table. The classical model (itself a family of slightly different proposals, but broadly consistent) gives you the familiar system with voiced stops, voiced aspirates, voiceless stops, and — depending on who you ask — voiceless aspirates. Then there's the glottalic model, independently developed in the early 1970s by Gamkrelidze & Ivanov and by Hopper, which reinterprets the whole stop system: the classical voiced stops become ejectives, the voiced aspirates become plain voiced stops, voiceless stops remain as such, while voiceless aspirates get demoted to allophones (whatever that means).

I wanted to build a toggle that lets the user switch between the two. My first instinct was simple: just swap the symbols and relabel the distinctive features and their values: *b becomes *pˀ, and voiced becomes ejective, done. No symbol should move to a different cell, and the table should not undergo a restructuring.

But it turns out the two systems are not congruent in that sense — they can't be converted into each other by a simple renaming of features.

Let's consider a fragment of the PIE phonological inventory:

Now let's try to switch the symbols and substitute the labels. Applying the substitution gives the following: p→p, b→pˀ, bʰ→b and Voice± → Ejective±, Aspirate± → Voice±. The outcome should be the following:

But wait, is the new *b ejective as well? It is not: in both versions of the theory, it represents the plain voiced series.

Then, perhaps, should the table have had the following form?

This solves the inconsistency, but violates my initial constraint that no symbol should move across the table. Indeed, the new *b replaces the classic *bʰ, and should remain in the same cell.

Try as you might, no rearrangement of cells and feature labels will make it work.

The truth is that my constraint is simply untenable: the glottalic theory assumes one ternary feature, rather than two binary ones. The only meaningful way of presenting the data is the following (even if it is far from being clear how such a three-valued feature should be named):

Note that the authors never mention this fact and, in general, pay little attention to the formal correctness of their phonological inventories (why am I not surprised?).

[artkeidan]

Siblings

All the teachers of English, all the time:

English is absolutely unique: it has the super-smart word sibling, which means ‘brother or sister’, regardless of gender. We do not have anything that useful in our language.

Tim Parks, Italian life:

He considered Verga’s story Black Bread. «I fratelli», the third paragraph begins, «che erano come le dita della stessa mano finché viveva il padre, ora dovevano pensare ciascuno ai casi propri». How difficult that was to translate! I fratelli, the brothers, but it actually includes the sisters too — the brothers and sisters. Italian has one word for both, a familiar, intimate word, quite different from the dry, technical sibling.

[artkeidan]

Interview: Slavonic fonts

A short interview on Slavonic scripts and fonts: https://type.today/en/journal/old-church-slavonic-today

[artkeidan]

Linguistics 101

Let me introduce you…

Picture taken in Portland, Oregon, around 10 years ago.

[artkeidan]

Deep embedding

I used to think structures like this existed only in the wildest fantasies of generativists. And yet, here is a real-life example of triple center embedding from Plato’s prose:

[artkeidan]

Lost in transcription

The following is planche no. 10 from the second volume of plates in the Encyclopédie by D’Alembert and Diderot (1763). Notice anything strange?

Here’s a hint: look at the Russian alphabet, specifically, the third column from the left. Where do those letter names come from?

The incredible answer is that they are actually letter names from the Georgian Mkhedruli alphabet, transcribed in Russian Cyrillic orthography.

This is almost mind-blowing, considering that the Russian and Georgian alphabets have virtually nothing in common; certainly not the letter names. Note that the Georgian alphabet is reproduced, with these same letter names, just a few pages later (plate 14).

Why would anyone assign Georgian names to Russian letters? Especially alongside the traditional Russian (i.e. Slavonic) names, which are even repeated three times — in the 2nd, 4th, and 5th columns of the plate.

There may have been some Georgian-speaking intermediary involved… but a quick search into the sources available to the Encyclopédie’s authors on the Russian language and writing system was inconclusive. I’m too lazy to dig deeper, so the mystery remains.

[artkeidan]

Grammatical genders according to Smotrytsky

Meletius Smotrytsky's Slavonic Grammar (1619) distinguishes the following values of the grammatical category of gender (the image below reproduces the relevant page):

мужеский 'masculine'

женский 'feminine'

средний 'neuter'

общий 'common'

всякий 'every'

недоуменный 'perplexing'

преобщий 'super-common'

In reality, there are of course only the usual three. Smotrytsky’s seven classes are inherited from the Latin tradition, but his literal translations of the Latin terms are particularly amusing in Russian.

[artkeidan]

The language of Adam and Eve

«Christian Europe used to consider Hebrew as the language of Adam and Eve, the first humans. But after the Middle Ages, many alternatives were suggested. This is an incomplete list of ‘first languages’»

See here: https://aeon.co/essays/why-is-linguistics-such-a-magnet-for-dilettantes-and-crackpots

Apparently, Dutch (with Flamish) wins the race.

[artkeidan]

Escher vs. Etruscan art

Tarquinia is an ancient city in the province of Viterbo (Italy). It is famous for its extensive Etruscan necropolises, over which the modern town is built, and for the unique and extraordinary painted tombs they contain. These tombs preserve magnificent and enigmatic frescoes, one of which, from the Tomb of Hunting and Fishing, is reproduced here:

M. C. Escher visited Tarquinia in 1927. Almost two decades later, he began to produce his well-known continuous patterns featuring animals and fish, such as the three reproduced here. I like to imagine that Escher may have drawn inspiration from the Etruscan memories of his youth when developing his own graphic style. In both cases, the birds are strikingly similar and fly in opposite directions (see the pictures below). Don’t you spot a similarity?

[artkeidan]

National deixis, or: when does “France” become “Orient” in translation?

The situation in which an author refers, in a text, to their own nation, country, language, or history can be considered an example of deictic (or indexical) word usage. Indeed, the notion of “my country” — not unlike the notion of “self” — shifts from one referent to another depending on who is speaking. Let us call this phenomenon national deixis. This kind of context often presents a challenge for translators and has been frequently discussed in the field of Translation Studies.

How should we translate a passage in which the author writes something like: “I love the hugeness of our country and the sound of its tonal language”? Most likely, we would translate the text word for word and then add something like “(the author is referring to China, — editor’s note).” The author knows what their country is, but the reader may not. The use of the possessive deictic pronoun our explains nothing on its own, since the referent of our, as uttered by the author, necessarily differs when it is rendered in the target language of the translation.

In the past, some translators have chosen to handle national-deictic references by substituting them with expressions referring to the nation, history, and language of the target-language speakers. Robert M. Adams, in Proteus, His Lies, His Truth (1972), famously warns against this approach. While translating a text from, say, French to English, he insists that:

“Paris cannot be London or New York, it must be Paris; our hero must be Pierre, not Peter; he must drink an aperitif, not a cocktail; smoke Gauloises, not Kents; and walk down the rue du Bac, not Back Street”.

I wonder, by the way, whether he was aware of a very similar passage in Eugène Ionesco’s absurdist comedy La Leçon (1951):

“Pour le mot Italie, en français nous avons le mot France qui en est la traduction exacte. Ma patrie est la France. Et France en oriental: Orient! Ma patrie est l’Orient. Et Orient en portugais: Portugal! L’expression orientale: ma patrie est l’Orient se traduit donc de cette façon en portugais: ma patrie est le Portugal". 

As is often the case, however, reality allows for some exceptions. I recently came across a compelling example of this same translational attitude which, far from being ridiculous, seemed entirely appropriate. In his book I Goti e la lingua gotica (1964), the Italian linguist Piergiuseppe Scardigli compares the remoteness of the Gothic kingdoms, from the perspective of 9th-century Carolingian scribes, to the equally remote memory of Spanish domination in Italy, from a modern Italian point of view:

“[…] nel IX secolo […] il regno di Teoderico era un ricordo tanto lontano quanto è per noi il ricordo della dominazione spagnola in Italia nel XVII secolo”.

This is a typical instance of national deixis: the author uses we to refer to the history of his own country. In the 1973 German translation of this book, however, the translator tacitly substituted “Italy” with “Germany,” and replaced the reference to the Spanish domination with the equally remote memory of the Thirty Years’ War:

“[…] im 9.–10. Jh. […] das Reich Theoderichs schon so lange der Vergangenheit angehörte wie für uns die Anwesenheit Gustav Adolfs auf deutschem Boden während des Dreißigjährigen Kriegs”.

[artkeidan]

Distanza lessicale tra lingue

Gira per internet un’accattivante infografica che presenta la cosiddetta distanza lessicale tra le varie lingue d’Europa. La metodologia non è specificata, ma viene fatto un vago riferimento all’articolo di F. Petroni e M. Serva, due fisici evidentemente appassionati di linguistica matematica ed NLP. Presumendo che l’infografica si basa su un qualche ragionamento analogo, propongo un po’ di pensieri critici.

La metodologia di Petroni e Serva si riduce, in ultima analisi, al calcolo della distanza di Levenshtein tra coppie di parole aventi lo stesso significato, provenienti da liste di 100 parole prese da ogni lingua in esame.

L'articolo contiene la seguente affermazione:

Indeed, our method for computing distances is a very simple operation, that does not need any specific linguistic knowledge and requires a minimum of computing time.

Che le conoscenze linguistiche non siano necessarie è, però, un'illusione (e non lo dico per difendere la categoria).

Un primo inghippo si nasconde nel concetto di "parole con lo stesso significato". In realtà, tali parole, cioè perfetti "translational equivalents" semplicemente non esistono. Ogni parola italiana viene tradotta non con una, ma con tante parole inglesi diverse, secondo il contesto, la sfumatura stilistica, ecc.; lo stesso capita per qualsiasi altra coppia di lingue. (Si veda un esempio banale: le possibili traduzioni del verbo italiano andare in inglese). Quindi la scelta di come riempire le liste di parole sulle quali si basa la misurazione è estremamente arbitraria e può trasformarsi in un bias capace di falsare il risultato. Ad esempio, in inglese esistono numerosi doppioni lessicali di tipo "Germanic vs. Latinate". Basta selezionare, per ogni significato, il termine "latinate", ed ecco che la distanza tra ENG e FRA diminuirà miracolosamente, ma indebitamente. Stessa operazione può essere fatta in molti altri casi (ad esempio, in russo esiste un vasto strato di doppioni lessicali di origine europea che affiancano parole di origine slava).

Secondo problema si nasconde nell'idea di confrontare la forma ortografica delle parole. Come sappiamo, le ortografie sono spesso arbitrarie, convenzionali e modificate a posteriori. Gli autori, probabilmente, avrebbero dovuto operare con una trascrizione fonetica normalizzata delle parole di ogni lingua, anziché con la forma ortografica: le due sono spesso talmente diverse l'una dall'altra da poter influenzare in modo significativo il calcolo finale.

Infine, la mancanza di “specific linguistic knowledge” impedisce agli autori di capire il fatto che questa loro "distanza lessicale" in realtà non è correlata con alcunché di scientificamente sensato. È come se dicessimo: classifichiamo le automobili per colore. Scopriremo magari che le macchine scure sono più performanti (perché, ad esempio, le macchine costose tedesche magari sono tendenzialmente di colore scuro). Però in realtà non è una misurazione scientificamente interessante. La distanza lessicale, analogamente, non misura un bel niente, perché gli autori non sanno cosa vanno cercando. Se si vuole ottenere una misura della distanza genetica, allora la "somiglianza" lessicale è un metodo totalmente sbagliato, che semmai impedisce di vedere le relazioni genetiche (le quali si basano sulle corrispondenze fonologiche regolari, non su somiglianze, come è noto a tutti i linguisti storici già da due secoli). Se invece si vuole misurare il numero di prestiti alloglotti in ogni data lingua, allora il calcolo è sbagliato in partenza, perché, appunto, non fa alcuna distinzione tra parole imparentate e quelle di prestito.

Come spesso accade in questi casi, la rappresentazione risultante, in ultima analisi, è molto poco informativa: non dice nulla che non sapessimo già (ossia che le lingue all'interno di un gruppo linguistico, ad esempio quello germanico, si assomigliano maggiormente che con quelle di altri gruppi). Va dato atto a Petroni e Serva: le loro conclusioni sono effettivamente molto caute.

[artkeidan]

When a genius gets insane…

The founder of modern linguistics, Ferdinand de Saussure, devoted the final years of his life to a rather peculiar research program concerning the presumed “anagrams” in ancient and, as shown here, even modern poetry in Indo-European languages. In this context, he contacted the Italian poet Giovanni Pascoli to ask whether certain Latin words might have been deliberately encoded as anagrams in his Latin poems.

…he is still a genius 

Saussure was fully aware that this line of inquiry was far from convincing and that it required solid empirical support. In this regard, it is worth looking at the final page of the letter, where he formulates a remarkably lucid observation, almost a methodological lesson, on the role of statistics in linguistic investigation:

“[…] plus le nombre des exemples devient considérable, plus il y a lieu de penser que c’est le jeu naturel des chances sur les 24 lettres de l’alphabet qui doit produire ces coïncidences quasi-régulièrement”

In other words, anagrams are likely to arise by chance, and the greater their number, the more chance itself must be regarded as the most plausible explanation. Precisely in order to rule out chance, Saussure therefore decided to address the question directly to the author, asking about his intentions.

Pascoli’s reply is unknown.

Saussure’s letters have been published online here.

Saussure’s research on anagrams has been collected and analysed here.

[artkeidan]

Two boys from Guinea...

In 1989 Abdoulaye Barry and Ibrahima Barry, two brothers from Guinea of 14 and 10 years of age respectively, asked their father why they could not write their language, and when their father wrote a few words in Fulani in Arabic script, they said “That’s not our script. We should have our own”. Then the two boys decided to devise a new alphabet for their language. Not unlike millions of boys do all over the world, at that age. With the difference that their creation — the Adlam alphabet — was successful. Over the next few years the Adlam script continued its development. It is currently in use in Guinea (where it is taught in schools), as well as in other countries, and its use is increasing. This is how the two brothers, now adults, tell the story themselves.

Here is a PDF with a handbook that teaches the script, in Fulani language.

(Note how the typical ABC pictures — mnemonically representing the letters by which the corresponding words begin — reflect a contemporary African reality.)