Jonathan Robie wrote: ↑May 12th, 2017, 9:55 am
Phonological features: Katharevousa contained various pronunciations that did not fit the Modern Greek phonological system. For example, νδρ (Ancient and Demotic /ndr/, katharevousa /nðr/); φθ (Ancient /pʰtʰ/), [α]υθ, [ε]υθ: Demotic /ft/, katharevousa /fθ/; σθ; ρθρ.
No reference is given for that claim ...
That statement is a little left of centre.
It is a mouthful for uneducated speakers of standard Modern Greek to pronounce those consonant clusters described as katharevousa in that statement. Older or more conservatively (right-wing = free-market) minded speakers, who had learned katharevousa would not have a problem with pronouncing them. In addition a few words with those clusters are in common usage - the nominative of "a man" formed by the addition of the sigma of the nominative to the third declension accusative form άντρας [ándras] coexists in the language with άνδρας [ánδras] - the implicit claim in that quote is that these coexisting forms are out of place.
The background to the sound differences in the diglossia is that during the Byzantine period the consonant clusters were simplified in various ways for example from two fricatives to fricative + stop. The ἀδελφός of Classical / Koine Greek has been "simplified" to αδερφός too.
The inaccuracy in the quote is that while it doesn't say, but means to say that σθ is simplifed to στ within the context of morphological changes from the Koine to Modern forms of the language. Morphology of the passive aside, -σθ- is still pronounced /sθ/ in words such as σφαγιασθείς [sfajiasθís], or μισθώνω [misθóno] (our μισθόω "to hire" - it is regular for Koine short vowel stem verbs in omicron to have the form ώνω if the survive into Modern Greek).