some of Ellis's central tenets - repeated and minimally amended

I should like to supply page numbers and other hints for my blog followers who are interested in knowing where I got my information from:
1. Social class divisions (vol. II, p. 629):

credit: National Portrait Gallery

Alexander John Ellis

1. 'Upper ten' (the court and nobles)
2. 'Middle class' (the professional and studious)
3. 'Commercial class' (the retail tradesman)
4. 'Young men and young ladies' (servants, porters, mechanics etc.)
5. 'Dangerous classes'

2. Ellis points to data collection problems:

1. He is aware of the observer paradox (vol. IV, p. 1086: "The only safe method is to listen to the natural speaking of some one who does not know that he is observed").
2. He is also well aware of the volatility of his data (vol. IV, p. 1086: "the sounds of language are very fleeting")

3. Data selection problem and insecurity:

1. Ellis points to the fact that there’s a great deal of variation as far as pronunciation of one and the same word is concerned (vol. II, p.628f.; last paragraph of p. 628 extending to p. 629). So he takes a descriptive stance (this is my personal conclusion; it shows his descriptive unwillingness to succumb to prescriptive absolutes.).
2. He introduces the conception of a mean incorporating variation (vol. I, p. 18: "there will be a kind of mean, the general utterance of the more thoughtful and more respected person of mature age, round which the other sounds seem to hover").

4. He sets up (at least) two categories of pronunciation:

1. received pronunciation (vol. I, p. 23: "recognize a received pronunciation")
2. dialectal pronunciation

5. Regional distribution of received pronunciation:

1. r.p. is regionally unmarked (vol. I, p. 23: "a received pronunciation all over the country, not widely differing in any particular locality").

6. Ellis's two caveats:

1. there's a certain degree of "regional colouring" of r.p (vol.I, p.23: "there will be a varied thread of provincial utterance running through the whole").
2. "received speech is altogether a made language, not a natural growth" (letter of 26 September 1882 by Ellis to Murray).