Feuerbach's Sentences – Translations compared by the Machine (Python Spacy)

Background

I taught German for philosophers in Italian universities for more than ten years. My students were acquainted with the History of Philosophy (they had done four hours weekly for three years at High School, because it has been considered a principal topic ever since the reform

done by philosopher Gentile in the Twenties) but knew only little German.

I often limited the reading material to just one sentence. This exercise gives little of what Markettari call contents, but quite an idea of how philosophers are thinking, i.e., their philosophical style.

Ludwig Feuerbach, Master of participle constructions

Participle constructions in German are developed recursively on the left, while Italian works on the right side. This means in German we can make use of a kind of parenthesis formed by the article and the noun and introduce very long constructions. In Italian you do not have

this possibility. In English you can do it, but usually you do not.

Feuerbach is famous for his mountainous sentences. He uses lots of participle constructions, typical for technical and administrative writing.

Hegel, Fichte even Kant had written in a different way. It could be interesting to see how the language structure in written texts changes during the nineteenth century. Especially academic texts are getting quite tiresome the last decades of the century,

A sentence and its translation

In one of those very long sentence in the Introduction to “The essence of Christianity” we find the following passage:

The book offers, Feuerbach says, the "Prinzip einer neuen, von der bisherigen Philosophie wesentlich unterschiednen, dem wahren, wirklichen, ganzen Wesen des Menschen entsprechenden, aber freilich gerade eben deswegen allen durch eine über-, d. h. widermenschliche, widernatürliche Religion und Spekulation verdorbenen und verkrüppelten Menschen widersprechenden Philosophie".

The students had fun searching for the noun „einer neuen“ does belong to. Needless to say, it is the last word, „Philosophie”.

With Python Spacy, we can analyze the dependencies in this sentence and design a dependancy tree.

The English translation by George Eliot gives:

"This philosophy is essentially distinguished from the systems hitherto prevalent, in that it corresponds to the real, complete nature of man; but for that very reason it is antagonistic to minds perverted and crippled by a superhuman, i.e., anti-human, anti-natural religion and

speculation". Sounds easier somehow. With diSplay: 

Instead of the nine levels found in the original, here we find only five. The structure is simpler. Maybe we could see the translation of a kind of Easy Reader. Nothing bad, as I myself have written lots of these books. But if we want to understand what is happening to philosophical thought during the nineteenth century, there might be missing something if we worked with translations. The structure, the way of thinking. 

Technically

Just a very small program for grammar parsing and tree displaying.

import spacy

from spacy import displacy

nlp = spacy.load("en_core_web_sm")

doc = nlp("This philosophy is essentially distinguished from ….")

for token in doc:

print(token.text, token.lemma_, token.pos_, token.tag_, token.dep_,

token.shape_, token.is_alpha, token.is_stop)

options = {"compact": True}

displacy.serve(doc, style ="dep", options=options, port = 5001)

The list for Feuerbach starts:

This this DET DT det Xxxx True True

philosophy philosophy NOUN NN nsubjpass xxxx True False

is be AUX VBZ auxpass xx True True

essentially essentially ADV RB advmod xxxx True False

distinguished distinguish VERB VBN ROOT xxxx True False

from from ADP IN prep xxxx True True

the the DET DT det xxx True True

systems system NOUN NNS compound xxxx True False

hitherto hitherto NOUN NNS pobj xxxx True False

prevalent prevalent ADJ JJ amod xxxx True False

, , PUNCT , punct , False False

in in ADP IN mark xx True True

that that SCONJ IN mark xxxx True True

it it PRON PRP nsubj xx True True

corresponds correspond VERB VBZ advcl xxxx True False

to to ADP IN prep xx True True

the the DET DT det xxx True True

real real ADJ JJ amod xxxx True False

, , PUNCT , punct , False False

complete complete ADJ JJ amod xxxx True False

nature nature NOUN NN pobj xxxx True False

of of ADP IN prep xx True True

man man NOUN NN pobj xxx True False

for toke")n in doc:

print(token.text, token.lemma_, token.pos_, token.tag_, token.dep_,

token.shape_, token.is_alpha, token.is_stop