An important topic that I have not found addressed almost anywhere. Regarding 'traditional music' ear-training, I have found it necessary to break the topic into temporal [ie rhythm], and non-temporal [sic] areas -- those focusing on aspects of pitch perception. I put melodic dictation largely into the non-temporal domain as the technique generally has been to notate the rhythm independently of pitch. The pitch can largely be converted to 'unmetered melody' to extract the sequence of pitches.
For pitched elements in tonality, I have broken this down into two large domains,  intervals, and  scale degrees. I have drilled them both independently, and moving forward at their own rates. [My reference here is to the two year ear-training course materials I wrote for an undergraduate music program.] For older students I have found that both need to be built on a solid basic music theory foundation. Generally, the student needs to be able to manipulate music theory symbols [intervals, triads, other chords, scales and some basic tonal harmonic functions] in order to progress reasonably easily.
In my reading through most of the commercially available ear-training texts, the underlying pedagogy is not explained. The composer / writer has made a number of [educated] assumptions about the sequence that materials need to be introduced in, and about how much drill is needed on these points. Hindemith is lean and clean, but as a sole source for most students is much too fast. Dannhauser is an excellent example of the progressive introduction of sight-singing elements, but few schools have the three or four years of courses to devote to the whole series.
Rather than the term 'perceptual block', I think of them as being 'learned patterns', and one way to unblock the perception, is to unlearn the pattern. In tonality, in scalar passages, thirds are 'equal' [sic]. The opening of the Beethoven Fifth has two descending thirds, which although of different absolute size -- major then minor, are equivalent at a structural level. ^5 > ^3 = ^4 > ^2, and the piece uses this structural identity. This movement uses the 'diatonic third' without differentiation compositionally -- the third is a distance in the scale, not an interval.
One way to 'hear' this intervalically to break down the scale degree recognition is to have it played backwards. .sdrawkcab deyalp ti si noitingocer eegred elacs eht cte. Or to play the two intervals as an atonal context: G G G Eb E E E C#. This leaves the interval intact while removing the tonal / scalar context.
2. My experience is that the 'correct' guide / anchor is required. There is a school of ear-training teaching that demonstrates intervals as being the first two notes in a song. The problem is that tonal hearing dominates, and the first interval of "The Star Spangled Banner" is not heard as a minor third, but as ^5 ^3 in the major scale.
From my reading, most ear-training materials do not come with a detailed, reasoned pedagogy. Many on-line ear-training 'courses' remind me of random-walk learning in that the drills have content, but little by way of form or context.
I think the literature you seek will not be found outside of pedagogical explanations that accompany specifically created / composed materials. It is one thing to draw up a chart of the sequence of [say] three years of dictation and sight-singing exercises -- with supplementary drill materials, it is, IMPE, another matter to produce the 48-60+ hours of dictation and 1,000 pages of sight-reading materials. Your kilometrage may vary significantly.
On 2012, Nov 25, at 1:24 PM, eldad tsabary wrote: