The Voynich Ninja

Minor update to the work with some small passages removed and some enlarged.

Question:
Is there a program with which I can translate the language in a PDF file?
Example English - German for PDF ?

There are two types of PDF: those from which the text can be extracted, and those from which it cannot.

In the first case, copy and past to Google Translate would work. Recent PDF's created from text editor files usually fall into this category.

In the second case (image data, or older text files), an OCR step would help. There are reasonable free on-line OCR sites available, but these tend to work from images rather than PDF.

I have used this about 2 years ago:
You are not allowed to view links. Register or Login to view.
but there may have been more developments in the meantime.

Thanks Rene
But I have to put the text into a translator.
What always annoys me is that Google-Chrome always translates the environment text but not the open PDF file.
It would be nice if he translates everything right away. Well, I can live with that, but I don't want.

New version of the work. Shorter pieces were removed, longer (at least paragraph) were kept.

(27-10-2020, 04:17 PM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.New version of the work. Shorter pieces were removed, longer (at least paragraph) were kept.

The Latin produced by this is ungrammatical and the translation is incompetent.

Let's look at the first Latin sentence: "Dēsōlō fōrmī aequus, aciē fōrmum. Assulā hastifōrma."

The latest version allegedly translates this as Giving up favorable forms, to sharp forms, to sharp-formed splinter.

Is the Latin good and the translation correct? As an initial matter, the translation isn't a sentence, but lets power on.

Dēsōlō means "I leave alone" or "I abandon/forsake/desert." In the middle ages it is used to mean "I ravage" or "lay waste" or "leave desolate". Itr can mean "I deprive" (if used with the ablative) or "I distress." In reference to a church, it means "I leave vacant." The reason I put in all the "I"s is that the verb is a first person present indicative. It is not a participle (or maybe a gerund), so "giving up" is wrong. To say that, one would need a participle "desolans" and a better, more idiomatic verb (perhaps "deficiens"?).

Dēsōlō is a transitive verb, and its object should be in the accusative case (or in the ablative with the meaning "I deprive"), but formi is in neither case. It's is either a genitive singular or a nominative plural, both options being ungrammatical. Plus, the word for "form" is a forma, a first-declension feminine, not a masculine. To mean "giving up forms" it needs to read something like "desolans formas". (It turns out there is a masculine formus, so that the "formi" is a valid word, but that word is an archaic adjective that means "warm.")

Turning to aequus, it is an adjective that means "even" or "level"; in some contexts it can mean "favorable" as in reference to a place or time. It also has to agree in case number and gender with its noun, but this nominative masculine singular, and neither formi nor formas are. It needs to read "aequas" to be in the same ballpark as the translation. Otherwise, it means something like, "I, a favorable person, abandon ..." (sorry, can't fit formi into the sentence).

So the first clause translates the ungrammatical Dēsōlō fōrmī aequus as if it said desolans formas aequas instead. Clearly the translation method is to look up meanings in the dictionary, ignore the grammar, and produce something that could make sense. This expedient is necessary because the Latin text is neither grammatical nor does it make sense.

Going on, aciē is not an adjective as the translation wants, but a noun in the ablative case. It means something like "by/with a sharp edge," though idiomatic Latin will have a preposition, which is missing. Indeed, none of the "Latin" texts in the paper have prepositions, which is a problem. To get the adjective "sharp" as the translation wants, "acras" is needed.

Formum does not mean "to ... forms"; the preposition is missing, and the noun is singular. It is also the wrong gender, and therefore means "a warm thing." "in formas" is better reflective of the translation.

Assulā is ablative singular, "by/with a splinter", which is the wrong case and wrong meaning.

[font=Tahoma, Verdana, Arial, sans-serif]hastifōrma is not the dictionary but appears to mean "spear-formed". It is also in the wrong case, here nominative.[/font]

[font=Tahoma, Verdana, Arial, sans-serif]I could go on. Every Latin sentence and its translation has similar problems. [/font]

Stephen Carlson, thank you for the reply. You are of course right in almost every aspect.

(13-11-2020, 03:56 AM)Stephen Carlson Wrote: You are not allowed to view links. Register or Login to view.Clearly the translation method is to look up meanings in the dictionary, ignore the grammar, and produce something that could make sense. This expedient is necessary because the Latin text is neither grammatical nor does it make sense.

....

none of the "Latin" texts in the paper have prepositions, which is a problem.

Many thanks to Stephen for going into the details of this passage!
What he writes of course applies to all Voynich "translations" so far. Since grammar is the weakest point in all proposals, I believe that the most efficient way to evaluate a solution is checking this aspect. Solvers utlimately deliver grammatical English, but is the grammar structure in the original Voynichese passage?

The conjunction is a trivial example: when a translation includes "and" does a corresponding word appear in the "word salad" generated by mapping Voynichese into words of the source language? Typically, word lists like "cat dog" are translated as "the cat and the dog", but this can only be legitimate if the proposed language had no conjunction.
As Stephen wrote, a similar argument can be made for prepositions: if prepositions appear in the English translation but are missing in the word-to-word mapping, the translation likely is the product of the solver's fantasy.

(13-11-2020, 08:59 AM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.Since grammar is the weakest point in all proposals, I believe that the most efficient way to evaluate a solution is checking this aspect

MarcoP, I completely disagree with that. Moreover, I believe that this sort of thinking prevents VMS from being decoded. But that is just a personal opinion, nothing more.

A general question is how would you evaluate a solution based on constructed language? What would be the technique there?

@farmerjohn, nothing personal, but I disagree.

(13-11-2020, 09:23 AM)farmerjohn Wrote: You are not allowed to view links. Register or Login to view.MarcoP, I completely disagree with that. Moreover, I believe that this sort of thinking prevents VMS from being decoded. But that is just a personal opinion, nothing more.

I would say that this sort of thinking prevents that incorrect VMS decodings can be accepted.

The discussion is often between people who know a great deal about a certain subject, and people who know less or little.

(Note that this is not just true for Voynich MS discussions).

A proposed translation in Latin, Hebrew, Nahuatl or any other language can only be judged properly by people who are intimately familiar with these languages. Not by someone relying on Google Translate.