Yes the word-for-word chart confirms the words themselves are Archi, not Cebuano or Quechua. But I finally think you're onto something. The vast variety in Spanish translations you are seeing mirrors the English (and Spanish) results Google is giving, and do not mirror an ordinary linguistic analysis of the submitted texts.
That means a possible explanation is something like: OpenNMT is being used both at Google Translate and at QuechuaTranslator, it originally used the Archi language as a convenient rarely used language for some kind of testing and tied it to various Bible texts with the randomness associated with machine learning, and then DigitalCoding used that fact to (wrongly) diagnose Archi as Quechua.
P to W is irrelevant because you get meaningful texts often when you change any single letter to another randomly. Try it.
Q's text, and my five sample texts in OP, at QuechuaTranslator (with Spanish Google Translate in parens) yield the following:
Gentileza de las cálculos catástrofes catástrofes catástrofes que aseguraron las bellezas mexicanas . (serán castigados y serán castigados por sus pecados.)
Respuestas prácticas bíblicas de bos bos bos bales (serán castigados)
Gentileza de pequeñuelos catástrofes catástrofes arrojados a las bellezas mexicanas (Que quieres hacer con tu vida?)
Gentileza de las catástrofes catástrofes catástrofes catástrofes que aseguraron las bellezas de robos mexicanos . (el torbellino y las águilas son los que abren la salida)
Gentileza de pequeñuelos cercanos australianos aproximados a la belleza bónica , comprendidos en bos bos bos (Te amo y te amo porque te amo.)
Gentileza de su abuelo acaudalada , acaudillado por los ejércitos afectados (que es varón de dolores, y experimentado en quebranto; y como que escondimos de él nuestro rostro;)
This suggests the two are not using the same algorithm, but that Google Spanish is yielding the same bizarre results as Google English, while QuechuaTranslator is yielding varied results (as if from machine learning) but with more regularity. Sometimes it appears the original Archi meanings (from the spreadsheet) are filtering in.
That means that the meaning of Q's statement would need to be limited to (1) the plain "translation" given by Google English/Spanish, (2) the potential encoding one can compile of Archi concepts nonnatively strung together, or (3) the fact that these translators are using a database that may contain other more interesting data. However, any of these other translations that you and I are uncovering are just part of the database and do not indicate they are Q's meaning as specifics.
Yes the word-for-word chart confirms the words themselves are Archi, not Cebuano or Quechua. But I finally think you're onto something. The vast variety in Spanish translations you are seeing mirrors the English (and Spanish) results Google is giving, and do not mirror an ordinary linguistic analysis of the submitted texts.
That means a possible explanation is something like: OpenNMT is being used both at Google Translate and at QuechuaTranslator, it originally used the Archi language as a convenient rarely used language for some kind of testing and tied it to various Bible texts with the randomness associated with machine learning, and then DigitalCoding used that fact to (wrongly) diagnose Archi as Quechua.
P to W is irrelevant because you get meaningful texts often when you change any single letter to another randomly. Try it.
Q's text, and my five sample texts in OP, at QuechuaTranslator (with Spanish Google Translate in parens) yield the following:
Gentileza de las cálculos catástrofes catástrofes catástrofes que aseguraron las bellezas mexicanas . (serán castigados y serán castigados por sus pecados.)
Respuestas prácticas bíblicas de bos bos bos bales (serán castigados)
Gentileza de pequeñuelos catástrofes catástrofes arrojados a las bellezas mexicanas (Que quieres hacer con tu vida?)
Gentileza de las catástrofes catástrofes catástrofes catástrofes que aseguraron las bellezas de robos mexicanos . (el torbellino y las águilas son los que abren la salida)
Gentileza de pequeñuelos cercanos australianos aproximados a la belleza bónica , comprendidos en bos bos bos (Te amo y te amo porque te amo.)
Gentileza de su abuelo acaudalada , acaudillado por los ejércitos afectados (que es varón de dolores, y experimentado en quebranto; y como que escondimos de él nuestro rostro;)
This suggests the two are not using the same algorithm, but that Google Spanish is yielding the same bizarre results as Google English, while QuechuaTranslator is yielding varied results (as if from machine learning) but with more regularity. Sometimes it appears the original Archi meanings (from the spreadsheet) are filtering in.
That means that the meaning of Q's statement would need to be limited to (1) the plain "translation" given by Google English/Spanish, (2) the potential encoding one can compile of Archi concepts nonnatively strung together, or (3) the fact that these translators are using a database that may contain other more interesting data. However, any of these other translations that you and I are uncovering are just part of the database and do not indicate they are Q's meaning as specifics.