http://www.dolmetsch.com/musictheory21.htm
In lesson 4, we discussed the way in which beats can vary in strength and how, through a suitable choice of time signature, the composer can make clear the rhythmic structure, formed by a particular pattern of strong, medium and weak beats. On many musical instruments there are limits to the strength of a beat, if all one can do is to blow, bang or scrape more or less enthusiastically. Wind-instruments blown hard play sharp and the tone is coarsened. The same instruments blown too gently will tend to play flat or not at all.
However, a performance in which all the notes are equally loud or soft is immeasurably duller than one where there is dynamic variety.
The dynamic detail can be in the note's strength when it starts, whether or not it is preceded or followed by a silence, the note's length and the mean strength of the note while it sounds. All these can be determined with suitable notation and we will look at each of these in turn.
Music, like written prose, tends to be made up of short sequences we call phrases. Consider Swift's 'A satirical Elegy on the Death of a late Famous General' (1772) - actually, on the death of Marlborough, the victor at Blenheim - in which each line is a single phrase.
But what of what, his friends may say,
He had those honours in his day,
True to his profit and his pride,
He made them weep before he dy'd.
Each line expresses a single idea which is the fundamental characteristic of a phrase. Of course, the choice of phrase length is not 'set in stone'. One might 'feel' that a more natural phrase length here is the pairing of lines (1 with 2, 3 with 4).
This freedom to feel poetry in various different ways occurs also in music and phrasing is a matter best left to the performer to communicate to the listener as he (or she) thinks best. We will return to this point when discussing the setting of words to music where the phrasing of the words tends to find its mirror in the shaping of the musical line.
The desire of editors and composers to make their intentions clear down to the very last detail includes the use of large sweeping 'slur-like' lines called 'phrase marks'. Slurs, which tend to embrace a smaller number of notes, help to shape the musical line even within broader phrasing marks and performers must be able to distinguish between them. On wind instruments, all the notes under the slur except for the first, are untongued, the breath flowing continuously while the fingers move. On stringed instruments, the equivalent effect is achieved by using a single sweep of the bow for each slur or phrase. On keyboard instruments the notes are played legato (smoothly) and with a light touch.
The slur removes the attack from the start of each note under it except for the first so providing a contrast in strength, a dynamic variety, between the first and the later notes. If slurring is to be effective, or indeed a distinction made between different phrases, the performer must avoid playing unslurred notes too smoothly. Evidence from the eighteenth century suggests that music then was played in a more detached manner than we associate, say, with the repertoire of the late-Romantic.
Daniel Gottlob Türk writing in his Clavierschule (1789) says that:
"...when notes are to be played in the usual manner, that is to say, neither staccato nor legato, the finger should be raised from the key a little earlier than the value of the note requires."
Reference:
Clavierschule, Daniel Gottlob Türk (1789) - translation of extracts, from German to French, by Jean-Pierre Coulon
By the nineteenth century Muzio Clementi (1801) offers slightly different advice, recommending a legato style.
"The best general rule is to keep down the keys of the instrument, the full length of every note."
Clementi's advice echoes the earlier concerns of François Couperin and J. S. Bach, for as Skip Sempé writes:
"Couperin wrote the Préludes of L'Art de Toucher le Clavecin as pieces to form the hand into making a certain very special sound - a sound he considered important from the very first contact with a harpsichord. Bach made the same point in his Inventions, where he refers to the central idea of cantabile art, that very quality which makes keyboard instruments so illusionistic: the magic fusion of a lyrical effect and its accompaniment achieved by the diversity of sound created by the art of touch."
Reference:
François Couperin: L'Art de toucher le clavecin
The advantages of a detached manner when playing in a large acoustically resonant building become clear. When the notes 'ring on' around the room, the harmonies overlap instead of flowing neatly one into the other. Slurring, in such surroundings, would obsure the line, and so the performer has to be able to adjust the performance to the demands made by the surroundings by ignoring slur and phrase marks that may, in this situation, have become redundant.
Slurs are distinguishable from ties, which we met in lesson 2, because ties only link together notes of identical pitch (e.g. B to B) while slurs never link together notes of identical pitch.
They pick up on this stuff, instantly; much more than dynamic drivers (that have other strengths). Go listen to a harpsichord or guitar for instance, played on an estat. It's a different experience. And as it's been mentioned before, PRAT is essentially referring to microdynamics or dynamic accents. 
