Mindful of the DonJuan narrator's injunction that "Good workmen never quarrel with their tools," I have in the main little other than praise for the second volume of Duncan Wu's fine project to survey in exacting detail William Wordsworth's reading.  Wordsworth's writing and now reading life continue to be extraordinarily well-served by the instruments, large and small, of scholarship: the Clarendon Press Letters (finished in 1988 and already followed by a 1993 Supplement volume, whereas the Coleridge letters rest unattended in book form since 1971), the Cornell Wordsworth volumes, the Owen and Smyser ProseWorks , Mark Reed's chronologies, the Shaver library catalogue, a linked series of bibliographic guides (Bauer, Logan, Stam, Kroeber and Jones), the venerable Lane Cooper concordance, Jared Curtis's 1993 edition of the Fenwick notes—all of these and more now available as print volumes even as schemes move forward for various electronic archives and utensils. Wu's two volumes assume on the reference shelf a more rigorous place beside what remains a useful ancestor, Markham Peacock's 1950 assembly of TheCriticalOpinionsofWilliamWordsworth ; the Cambridge University Press is to be commended for continuing to devote resources to this essential branch of scholarship while it also offers at the front of its Romantics list the volumes of the Cambridge Studies in Romanticism series. 
The useful traffic between these two types of projects may be simply illustrated: students alert to the work to theorize gender and romantic writing can now locate in Wu's volumes the precise evidence (through 1815) of the extent of Wordsworth's engagement with the writing of, simply among contemporaries, Aikin, Baillie, Barbauld, Edgeworth, Lamb, Lickbarrow, More, Opie, Robinson, Seward, Smith (Elizabeth), Smith (Charlotte), Williams, Wollstonecraft, and others. The volumes not only document discrete events in their alphabetic march (in this volume, from a pseudonymous author named "A" to "Edward Young"; Wordsworth appears never to have read—through 1815—an author or book whose name or title begins with Z, a fact hitherto invisible) but also yield a cumulative significance in many larger categories of reading, such as the Greek and Roman classics, medieval and renaissance writing, continental writing, periodical publishing, journalism, and the contemporary novel. I find most fascinating Wu's indefatigable work to reconstruct the culture of libraries known to Wordsworth—private, semi-private, institutional—in late 18th and early 19th century Britain, a rich body of generally untapped evidence the freshness of which demonstrates the necessary continuity between the kind of work that underwrites what the profession used to call philology and what it now prefers to honor as cultural studies. As these volumes document Wordsworth's immersion in many different areas of writing, publishing, book collection, and book circulation, to my mind the oddest category of reading that Wu has set himself the task to track is what the latter-day caricaturists of Wordsworth the egotist might maliciously predict would be the only category of a book titled Wordsworth'sReading : Wu dutifully documents when possible Wordsworth's reading of his own works, a task the impossible fullness of which might be considered roughly analogous to trying to infer from a few known visits how often Michelangelo slipped back in the Sistine Chapel and looked up. That some information is more useful than other bits in no way diminishes, however, the bedrock principle that I would here uphold, and that Wu's volumes superbly illustrate: the more information, the better—so let's have that curious record of when Wordsworth read himself (when didn't he?) and let students make of it what they will.
A rigorous evaluation of Wu's methods in shaping his often hard-won information is already available in Bruce Graver's 1995 review of the first volume (1993) of Wordsworth'sReading , and to do much grumbling back down the trail in the wake of these two sharp archival sleuths would only set me up for Hamlet's self-admonishment in the graveyard: "The hand of little employment hath the daintier sense."  Before turning, then, to what I might more usefully contribute to this conversation, a set of concluding comments about the interpretive uses to which this body of evidence is put, I will discharge the obligation to demonstrate why I have any claim to be out on the trail myself in the form of two quick examples of instances in which Wu's feast of information needs to be supplemented—a fact that Wu is himself of course the first to declare ("A work of this kind is by its nature incomplete") and that his admirable labors arguably should give him exclusive rights to declare, were life unencumbered by reviewers.
In the early autumn of 1820 in Paris, Wordsworth finally met Helen Maria Williams, to whom he had carried a letter of introduction (in the event, unused) from Charlotte Smith in 1791. In a footnote to the sonnet "To Hope" in Williams's 1823 PoemsonVariousSubjects , we learn that "I commence the Sonnets with that to HOPE, from a predilection in its favour, for which I have a proud reason: it is that of Mr. Wordsworth, who lately honoured me with his visits while at Paris, having repeated it to me from memory, after a lapse of many years."  Williams's "To Hope" first appeared in 1790 at the end of chapter XVI of the first volume of Williams's novel Julia , and then reappeared in the expanded 1791 edition of her Poems , which suggests that either or both of these books need to be added to the list of the other Williams volumes that currently appear in Wu's lists—or we need to determine where else Wordsworth encountered the text of "To Hope" that he recited from memory in 1820. A second brief example: When tracking Wordsworth's reading of Coleridge, Wu notes that Wordsworth's reading of the earliest version of what became the "Dejection" verses on April 21, 1802 was "the beginning of the poetic dialogue that continued with ResolutionandIndependence , Dejection : AnOde , and W's Ode ." If we take the word "dialogue" here to signify among other topics and tropes the spousal, then the text that's missing from this list is the "Prospectus" to The Recluse , and Wu's statement that Wordsworth's reading of Coleridge's earliest "Dejection" drafts "begins" this dialogue needs to address the arguments Beth Darlington summarizes in her 1977 Cornell edition of HomeatGrasmere , that, in the earliest "Dejection" drafts, Coleridge is in fact not initiating this dialogue but responding to Wordsworth's prior figures in an even earlier version of what became the "Prospectus" verses.  To compound such examples would only draw out at unnecessary and unseemly length a principle that I fear is not as obvious as it once might have been: as is the case with all scholarly tools, even (especially) the apparently magisterial, Wu's volumes should be used not only with much gratitude but also with much care.
The wealth of data in these volumes comes packaged with relatively modest claims about its argumentative ends. In the very brief "Preface" to the first volume, Wu announces that the project emerged from "the desire to offer a corrective to the various misleading remarks made both by Wordsworth's contemporaries and by the poet himself about his reading." The second volume has a longer preface in which Wu does more work to locate his labors historically, citing as precedent Lane Cooper's 1907 ModernLanguageNotes essay on Wordsworth's reading that Graver brought to attention in his review of Wu's first volume. The language of Wu's second preface does more often gesture toward the larger interpretive ends to which the data might be put: it speaks of how the first volume "chronologized [Wordsworth's] intellectual life" and invites readers to supply information "to provide an even fuller account of Wordsworth's intellectual life."  These phrases open up a host of interesting questions, including whether Wordsworth's intellectual life may be said to be coextensive with his reading. To round off this assessment, I would like to turn at the last to address some of the interpretive assumptions upon which the two volumes appear to me to rest.
Wu's books are shaped to adjudicate on evidentiary grounds an old and apparently modest philological question: what did Wordsworth read? But because that question has, from De Quincey onward, an historically entrenched answer ("Not much"), and because that entrenched answer has been assumed to be in evidence in very large arguments about how we read romanticism (most recently, as corroborating evidence to discount the value of the egotistical sublime), Wu's project carries an understated polemical edge. Because he offers a new answer to the old question ("Much more than we used to think"), Wu's new answer may be read to imply a response to the conclusions that assume the old answer as evidence: Wordsworth in these volumes is shaped to appear much more of a communitarian (on paper, as it were) than some would give him credit for, an eagle on the wing flipping pages (if such a thing is aerodynamically possible) rather than an eagle in a nest unfurnished with books, to run with Keats's figure.
Both sides of this old argument about Wordsworth as it here centers on the issue of Wordsworth's intellectual life strike me as resting on a questionable assumption about reading, which we might call the Attenborough Thesis. In his 1993 film Shadowlands , Richard Attenborough puts in the mouth of C. S. Lewis a sentimental mantra (which Lewis lifts from a kleptomaniac student, who has pinched it from his schoolmaster dad): "We read to know we are not alone." I produce this text to point out that there is at least one other question at issue in this debate in addition to what Wordsworth read (or didn't read). That question, of course, is Why did Wordsworth read—or not read? That question is not explicitly addressed in Wu's volumes (or rarely elsewhere in the debates on Wordsworth's reading) because, it seems to me, both sides of the debate assume that something like the Attenborough Thesis is the answer to the question. The unstated argument seems to run along these lines: the more one reads, the more one is open to the other; the less one reads, the less so; more is better. I want to suggest that such a limited set of assumptions is inadequate to the complex traffic between books and life. Even the film itself unintentionally gives the lie to the sweetly intended truth of the thesis: are we to conclude that reading prepared a bookish don to meet Debra Winger (should we applaud the fact that he was looking for love, but in all the wrong places?) or are we to conclude that reading in some way blocked or delayed that meeting (should we lament the fact that he was looking for love in all the wrong places?)? Is reading enabling or disabling? Or how is it both?
Consider, for the sake of argument, a set of contrary forms of the Attenborough Thesis ("We read to know we are not alone"), a form of the argument that reading enables the social:
We read to know we are alone (in which form reading celebrates the absence of the social).
We read not to know we are not alone (in which form reading blocks or disables the social).
We do not read to know we are not alone (in which form avoiding reading enables the social).
We do not read not to know we are alone (in which triply-negated form avoiding reading avoids the absence of the social).
This last formulation seems to me to begin to get at something important going on in Wordsworth's reading. It might be paraphrased as "We don't read to avoid confronting just how alone we are," an idea that would suggest how strange, foreign, and intractably other the other is, how risky that encounter is, how dangerous it is. The other's discourse might take us over and inhabit us—not a colonization of other by the self, but the opposite, what Gerald Bruns is driving at in this passage from his 1989 essay "Wordsworth at the Limits of Romantic Hermeneutics":
We should think of Wordsworth as . . . giving a critique of the subject by dramatizing its vulnerability to otherness, the abysmal risks that hover and loom both inside and out. I read Wordsworth's celebrated and very guarded approach to the reading of books as being of a piece with his desire to keep his distance from others, to preserve self-possession against the demonic character of another's discourse. Discourse is demonic, possessive, dispossessing. It is perhaps the demonic nature of discourse that gives us the idea that we have consciousness, feelings, imagination, a poet's mind. 
In this line of thinking, how we interpret quantity of reading (more is necessarily better) can be crucially shifted; a comparative paucity of reading may be explained not by a lack of sympathy (as in the conventional view) but instead by the risks of an excess of sympathy, a too-total immersion in the other's discourse to the extent of loss of self.
In order better to understand Wordsworth's reading, I would want in addition to Wu's excellent volumes a fantasy companion archive of what he didn't read: what Wordsworth had the opportunity to read and chose not to, or where reading broke off, or an account of writers that he read something of and then later chose not to return to. In the company of just the great romantic readers (and from just the Wordsworth circle we could cite Southey, Coleridge, Crabb Robinson), Wordsworth plainly read a good bit less—a great deal more than we previously thought, thanks to Wu's labors, but still, on the whole, less. But it must be stressed that that fact can be interpreted in several different ways, and I am inclined to resist the prevailing assumption from De Quincey onward that it is necessarily a discreditable fact. Some of the great readers of history surely also have been, paradoxically, some of the great nonreaders, a class of readers who sometimes must have chosen not to read because they had discovered just how dangerous it can be. In order to ponder these questions, in order to begin to arrive at that "even fuller account of Wordsworth's intellectual life" that Wu invites, there is much reading to be done, the remarkable extent of a great deal of which Duncan Wu has skillfully mapped for us.
The first volume is Duncan Wu, Wordsworth's Reading 1770-1799 (Cambridge: Cambridge Univ. Press, 1993).
Markham L. Peacock, Jr., The Critical Opinions of William Wordsworth (Baltimore: The Johns Hopkins Press, 1950). Peacock is still valuable as a guide to the post-1815 material, and Wu in at least once instance (p. 261; the issue—not inconsiderable—is Wordsworth's reading of Kant) relies on Peacock's report of evidence.
The review appears in The Charles Lamb Bulletin, New Series 91 (July 1995) pp. 164-67; see also Bruce Graver's 'Duncan Wu's Wordsworth's Reading: 1770-1799: A Supplementary List with Corrections,' Romanticism on the Net 1 (1996) <http://users.ox.ac.uk/~scat0385/wordsworth.html>.
Helen Maria Williams, Poems on Various Subjects (London, 1823) p. 203.
William Wordsworth, Home at Grasmere, ed. Beth Darlington (Ithaca: Cornell University Press, 1977) pp. 21-22; Darlington here summarizes with permission from "his manuscript notes" the case for dating the "Prospectus" manuscripts worked out by Mark Reed subsequent to the second volume of his chronology in 1975. Paul Magnuson notes and builds on this argument in Coleridge and Wordsworth: A Lyrical Dialogue (Princeton: Princeton University Press, 1988) pp. 305-06; see also Gene W. Ruoff, Wordsworth and Coleridge: The Making of the Major Lyrics, 1802-04 (New Brunswick: Rutgers University Press, 1989) pp. 100-01.
The chronological focus announced by the two volume's titles is in some fundamental respects misleading; because of the books' alphabetic organization, they do not in practice facilitate chronological questions. If you want to know what Wordsworth was reading in the summer of 1809, for example, you'll need to work your way through the whole book from A to Z, watching for that date. Wu states in the first volume that it is intended to serve as a "companion" (I.ix) to Mark Reed's chronology of the same range of dates; in the second volume, the relationship has taken a Blakean turn, as volume two now "shadows" (II.ix) Reed's second volume, of the same dates; but these statements, like the books' titles, promise a chronological utility that simply isn't there. Pondering this question, it has also occurred to me that the dates of the title of volume one, if taken literally (and how else can we take them?) put forward a remarkable claim (Wordsworth's Reading 1770-1799); I haven't combed the volume to see if there really are entries from 1770 and 1771, but if there are, I think that's news.
Gerald L. Bruns, "Wordsworth at the Limits of Romantic Hermeneutics," Centennial Review 33 (1989): 415.