448 OBSERVATIONS ON THE LANGUAGE OF CHAUCER. 
and the comparative value of the manuscripts he employed, but the grammatical 
rules he has given us are both inadequate and inaccurate, while he puts at the 
head of the five manuscripts to which he ascribes most credit a very “ incorrect 
and carelessly written volume,” * part vellum, part paper (Harl MS. 7335), and 
excludes from this list the very best manuscript in the Museum (Harleian 7334). 
We do not know on what principles the order of the manuscripts used was settled, 
but correct philological principles were certainly not the guide. One manuscript was ! 
taken as a standard for a time, then another, and then a third. The impropriety of 
such a procedure is obvious on a moment's reflection, and will be forcibly felt, if 
what has been said of the unsettled state of English at the end of the fourteenth cen- 
tury, and of the liberties taken by copyists, is borne in mind. The natural result of 
an arbitrary compounding of a dozen manuscripts, representing the dialects of various 
dates and localities, (not without an admixture of the idiosyncrasies of their respec- 
tive writers,) is an artificial text, conformable to the actual speech of no time, place, 
tribe, or individual. And this, so far as we can see, was the process of Tyrwhitt.j 
The Harleian manuscript No. 1334 was made the basis of a new edition of the 
Canterbury Tales, prepared by Mr. Wright for the Percy Society (1847-51). This 
manuscript was “collated throughout” with the Lansdowne MS. 851, in the British 
Museum (which seems to be Tyrwhitt’s W), and as far as the Wife of Bath’s 
Tale with two others. The collations, however, do not extend to grammatical 
minutie, and though the editor informs us that he has corrected many obvious 
errors, we may regard the text as essentially a reprint of the Harleian MS. 7334. 
As such it is of great value, but it is, nevertheless, by no means a satisfactory, or 
even a comfortably readable text. The number of manifest errors still left is 
considerable, the number of probable ones enormous. Hundreds of lines are in- 
complete, and long passages exhibit much irregularity of language and metre. On 
* These are the words of Sir Frederic Madden, in answer to an inquiry of mine. 
T Through the kindness of my friend, H. T. Parker, Esq. of London, a zealous lover of Chaucer, I have 
in my possession Tyrwhitt's original collations of nearly all the MSS. in the Museum and at Cambridge men- 
tioned in his list. The Oxford collations are not included. Tyrwhitt took an old printed copy, and corrected 
it minutely on the margins according to the various MSS. which he adopted as authoritative, sometimes chang- 
ing at an interval of less than a page, sometimes keeping on with one and the same for ten pages. He then 
entered the various readings of other MSS. on blank leaves. Those marked in his list A C, C,, T, W, are 
most used. The marginal corrections are, more than half the time, made according to A, the inferior codex 
spoken of above; about one third of the time according to C,. C (the excellent Harleian MS. 7334) is the 
guide for only the first two pages of the Prologue, but is collated throughout. For exactness’ sake, I mention 
that the Persones Tale is wanting in these collations, except the very beginning, 
t 
