Internet Shakespeare Editions

Blindness and insight: Damaged type, damaged eyes, and Q2 Hamlet

Eric Rasmussen

This article originally appeared on the site of ArdenOnLine

A team of graduate students and I have completed an extended type-recurrence study using four of the seven extant copies of Q2 Hamlet. (1604). Our preliminary findings suggest that the printer's copy for Q2 was cast-off for setting by formes. This exciting discovery, if confirmed, will have a significant impact on our understanding of the nature of the manuscript behind this important Shakespearean text. Specifically, the conventional belief that the printer's copy for Q2 was an extraordinarily difficult and variable set of Shakespeare's foul papers will have to be revisited if, as it now appears, the manuscript was, in fact, orderly enough for the compositors to make the fairly exacting calculations required to estimate precisely how much copy would be needed to fill each of the fifteen quarto gatherings (by counting words and making computations according to the sizes of type and page) before type-setting even began.

We are presenting our findings here on the Internet Shakespeare Editions site in the hope that other investigators in the global community of Shakespeareans with access to other original copies of Q2 will take the opportunity to cross-check our results and help us to refine our conclusions.

In no other field of scholarly endeavor do the practitioners complain as loudly about what they do as those of us who examine damaged type in order to reconstruct the printing history of Shakespeare's texts. Antony Hammond once characterized this "eye-breaking work" in which "frustration sets in on a major scale" as "the most time-consuming and potentially frustrating bibliographical labor yet conceived."[1] Given that type-recurrence investigators often sustain permanent eye damage in the course of their research, it's not surprising that we're rather myopic, particularly in our fascination with determining whether early texts were set seriatim or whether the printer's copy was cast off for setting by formes. The only satisfactory way to establish the order of composition is to investigate the recurrence of individually recognizable types.

Supported by a generous NEH grant * and armed with a set of sophisticated optical lenses, Robert Lerner, Marsha Urban, and I independently examined the British Library copy of Q2 Hamlet, type by type, under high magnification. Each type that differed in any significant respect from the usual appearance of that character was noted. In those cases in which we found sufficiently distinctive damage to allow for accurate identification of individual types, we made a sketch of each, recording the position of its initial appearance along with any subsequent appearances. Jennifer Forsyth simultaneously examined the Capell copy (Trinity College, Cambridge). The investigation proceeded at the rate of about one quarto sheet (8 pages) per day, and the team met every evening to compare notes. Lerner and I then repeated the investigation using the Huntington Library copy.

Despite the amount of attention that has been focused on the early texts of Hamlet, it is not widely known that Q2 was printed with two distinct fonts of pica roman type and two distinct fonts of italic.[2] The roman font used by Compositor X to set sheets B-D, F, I, and N-O is slightly larger (a capital "W" measures 4.5 mm in width) than the font used by Compositor Y to set sheets E, G-H, and K-M (in which a capital "W" is 4 mm wide); X's italic font is slightly smaller (Exeunt measures 9.5 mm) than that used by Y (in which Exeunt is 10.3 mm wide). Remarkably, the types did not at any time become mixed. Apparently, X always distributed the type from the pages he had set back into his case and Y distributed his pages into his case. It would appear that the supplies of type were relatively small, forcing each compositor to distribute part of the sheet that had just been machined before going on to set the next sheet: types from sheet B reappear in sheet C, types from C in D, etc. Interestingly, it appears that the outer forme was generally distributed before the inner forme. The outer forme of sheet G, for instance, contains types last seen in the outer form of sheet E, but no types from the inner forme of E; the inner forme of G contains types from both the inner and outer formes of E. This suggests that the outer forme of sheet E was distributed before outer G was set, after which inner E was distributed and inner G was set. This pattern is repeated throughout. Thus, it would appear that Q2 was not set seriatim but by formes. That is, the compositor would first set pages 1r, 2v, 3r, and 4v (the outer forme) before turning to 1v, 2r, 3v, and 4r (the inner forme). This hypothesis, which is corroborated by our analysis of recurring headlines, helps to explain certain irregularities in the Q2 text: F1r, F2v, and F3r, for instance, all have 39 lines per page, whereas the rest of the pages in the sheet contain 40 lines. So too, F1r, F2v, and F3r exhibit a different manner of prose indentation (what we would call a hanging indent) than that used elsewhere in the sheet. It would seem reasonable to conclude that pages F1r, F2v, and F3r were set first, after which the compositor seems to have changed his habits regarding the number of lines per page and the style for indenting prose.

Anyone who has undertaken a type-recurrence study knows that after the initial data is gathered a process of protracted subtraction begins. For instance, when Peter Blayney compared the damaged types he had identified in the Bodleian copy of Lear with those in the Trinity College copy, he found that more than a thousand supposedly damaged types were, in fact, simply inking flaws; and when subsequent scholars examined other copies of Lear , they challenged nearly half of Blayney's results.[3] The following table lists all recurrent types we identified in Q2 Hamlet. All lines on the page are counted (except the headline and blank lines). The word in which the damaged type appears is given parenthetically. A query following the listing indicates a probable identification, but one about which we remain uncertain after examination of multiple copies. We have tried to exert reasonable care to ensure that these listings are accurate. However, we acknowledge that our eyes no longer focus as well as they did before this investigation began, and we look forward to having our errors and oversights brought to our attention.[4]

Recurring types in Q2 Hamlet

Roman Font I

Used by Compositor X for sheets B - C - D - F - I - N - O

A1 B1v8 (And) D2r13 (And) F3v16 (Anon)
H1 N4v33 (Haue) O1r13 (Hath)
I1 F2v21 (I) I2v12 (If)
M1 F2r34 (Maiestie) I1r38 (My)
M2 F1r4 (Marke) I1v1 (May) N2r35 (Maiestie)
O1 B3r9 (O) C3v35 (O)
W1 D2v13 (Would) F4v3 (Weele) I1r6 (Which)
W2 B3r35 (Wherein) D2v13 (Would) F4v3 (Weele) I1r6 (Which) N2v38 (What)
a1 B1v9 (dreaded) C2v6 (made)
a2 I2r9 (haue) N3v16 (as)
b1 F1v33 (but) I2r25 (brother's)
d1 C2r7 (had) I3v12 (and)
d2 I4v20 (and) O1r27 (dead)
d3 N3v17 (readiness) O1v19 (dooes)
e1 B2r10 (selfe) C4r20 (lender)
e2 F4r13 (rehume) I2r18 (father)
e3 N3r12 (answere) O1v4 (Absent)
g1 F4r8 (Iigge) I1v32 (passage)
g2 F3r2 (light) I2v4 (King)
g3 N1r12 (Dogge) O1r30 (sergeant)
h1 B1v10 (haue) C1v2 (this)
h2 C1r7 (that) I3r7 (heydey)
h3 F4r13 (that) I1v29 (thought)
h4 C1r38 (That) I2v7 (thee) N4v11 (him)
h5 F4r32 (while) I3v3 (the)
h6 C1r7 (that) I3r7 (heyday)
i1 B1v16 (againe) F2r34 (Maiestie)
k1 I3r35 (speake) N4v10 (drinke)
m1 D3v14 (remember) F2r26 (animales)
n1 F3v19 (vnequall) I1v9 (then)
n2 I1v9 (then) N2r11 (stand)
o1 B1r7 (long) C1r15 (to) D2r12 (lookes) F1v34? (to) N2r13 (election)
o2 B3r3 (sound) C1r35 (How)
o3 C1r18 (comfort) F4v10 (Good)
o4 D1r13 (to) F1r26 (conception) I3r26 (reason)
o5 F1r23 (for) I3r33 (corruption)
o6 I3r1 (followes) N3r17 (not)
p1 B2v14 (appeare) F1v29 (priuates) I4r16 (Repent)
r1 B2v28 (graues) C3r21 (necessaries) D4v5 (Sweare) I3r2 (eare) N4r14 (for)
s1 C3v22 (gaulis) I1v25 (this)
s2 F4r18 (treason) I1v29 (course)
t1 N3r10 (triall) O2r29 (the)
ss1 C1r39 (Possesse) I4r26 (Assume)
st1 C4v14 (sterling) I3v3 (stole)
st2 F4r18 (against) I3r28 (turnst)
st3 B2r32 (last) C3v25 (blastments)
st4 I4v32 (trust) N3r29 (most)
th1 D4v19 (Earth) F3v24 (with)

Roman Font II

Used by Compositor Y for sheets E -G - H - K - L - M

A2 E2v14 (And) K4v6 (And)
B1 E3r16 (But) G2v11 (Be) H2r22 (But)
B2 G3v22 (But) K1v4 (Both)
D1 L2r10 (Doue) M4v13 (Dane)
F1 G4r9 (For) H2v3 (For)
I1 G3r5 (I) L3r34 (I)
M3 H2v25 (Madam) K1r16 (Mad) L3r23 (my)
S1 H4r4 (Sir) K4r2 (She)
a3 E3r32 (haue) L2v5 (and)
a4 E3r38 (lunacie) L2v14 (as)
a5 G1r6 (That) L3r5 (way)
a6 G2v10 (faire) H4v12 (unnatural) K1r5 (prating)
a7 L2r25 (all) M3v17 (alas)
b2 G4v3 (but) H4v34 (but) K1v34 (be)
d4 E2v31 (death) G3r20 (made2)
d5 G1r25 (dame) L4r5 (deuise)
d6 E1v37 (windless) G3v32 (pronoun'd) L1v28 (good)
d7 H1v19 (tragedy) K1v38 (body)
e4 E2r10 (farewell) H4r3 (friend) M4v19 (sunder)
e5 G1v16 (matter) L1v32 (like)
e6 G3v26 (conference) M3r5 (indeede)
f1 E2v32 (himself) G3r30 (forme) M1v5 (of)
f2 E4v4 (of) L2r20 (for) M4r26 (of)
f3 L4r1 (for) M2v1 (fellowe)
h7 E2r34 (that) G3v30 (vnmatched) K1v23 (thence) M1r1 (the) M4r19? (charitable)
h8 E4v37 (further) G1r14 (haue) H4v18 (with)
h9 H2v20 (heere) L4r13 (vnworthiest) M3v6 (why)
h10 G2r8 (doth) L3v31 (otherwise)
h11 G3r35 (harth) L1v27 (then)
i2 E1v13 (flight) G3v8 (determination)
i3 G4r31 (will) L1r37 (in) M3v26 (thing)
k2 G1r2 (kytes) M2r9 (folke)
k3 G1r8 (vnpacke) H3r6 (croking) L1v24 (soopstake)
k4 L3v18 (asking) M1r7 (mountibank)
l1 E2r7 (let) M4r4 (world)
m2 E2r4 (my) M3r11 (man)
m3 G2v11 (remembered) H4v12 (me) K4v16 (me)
m4 G3r10 (my) L2r9 (many)
n3 E2v23 (long) G4v15 (finger)
o7 G1v35 (hope) K2r36 (doost)
o8 E2r10 (Lord) G2r22 (consumation) M4v22 (Good)
o9 G4r29 (Lord) H4r12 (loue)
o10 H3v17 (from) K2r30 (for) K2r23 (some)
o11 L2r25 (Robin) M4v30 (God)
p2 G3r29 (expectations) H3r14 (poyson)
s3 G3r34 (see) L4r26 (so) M3r21 (so)
s4 L2r7 (should) M2r20 (shypwright)
s5 K3v5 (vs) M2v2 (busines)
u1 E3r14 (pleasures) G2v33 (beautie) L4v14 (begunne)
v1 H4v13 (vse) K1v31 (vp)
w1 H1r35 (two) K1v27 (owne) M2r21 (gallows)
y1 E4r33 (lyer) G4r22 (Player)
y2 G4v38 (they) H2v31 (tropically)


ct1 G3r29 (spectation) M1v34 (act2)
ct2 E1r28 (Adicted) G3v2 (affections) H2r9? (functions) L1v4 (Acts) M1v35 (act)
sh1 G4r3 (show) H2r10 (shalt) M2r8? (she)


E 1 G1v22 (Exeunt) H1v25 (Enter)
Q 1 E3v6 (Quee) G1V32 (Quee) H1v2 (Queen) K4r32 (Quee) L1r31 (Queen) M1r32 (Quee)
Q 2 H1v25 (Queen) M1r30 (Quee)
a 1/ H1v11 (Mallico) K2v38 (England) L1v19 (Laer.) M3r31 (Ham.)
a2 H3v21 (Ham) K1v27 (Ham)
m 1 H1r22 (Ham) M3r37 (Ham)
m 2 H3v21 (Ham)<</td> K1v27 (Ham)


* The New Variorum Hamlet project, co-edited by Bernice W. Kliman (project co-ordinator), Hardin Aasand, F. Nick Clary, and myself, is supported by a grant from the National Endowment for the Humanities, an independent federal agency. [Back]

[1] "The White Devil in Nicholas Okes's Shop." Studies in Bibliography39 (1986): 135-72, esp. 160 and 170. [Back]

[2] Although W. W. Greg noticed that Q7 Hamlet (1676) had been printed with two different types, he failed to observe this feature of Q2 (A Bibliography of the English Printed Drama to the Restoration [London: Bibliographical Society, 1939-52], I:312-13). In the 1955 issue of Studies in Bibliography , devoted largely to the texts of Hamlet, both John Russell Brown and Fredson Bowers refer to J. Gerritsen's discovery that two fonts of italic were used for Q2 (volume 7, pp. 25 and 45). The existence of two distinct roman fonts was first pointed out by W. Craig Ferguson in his study of Pica Roman Type in Elizabethan England (London: Scolar Press, 1989), p. 15. See also Adrian Weiss, "Bibliographical Methods for Identifying Unknown Printers in Elizabethan/Jacobean Books," Studies in Bibliography 44 (1991): 183-228, esp. 224. [Back]

[3] Peter W. M. Blayney, The Texts of 'King Lear' and Their Origins. Vol. 1. Nicholas Okes and the First Quarto (Cambridge: Cambridge University Press, 1982), p. 95; Antony Hammond, review of Blayney in The Library, 6th ser. VI (1984): 91-2; Paul Werstine, review of Blayney in Shakespeare Quarterly 36 (1985): 121-5. [Back]

[4] Although we employed the traditional technique of looking at original texts under high magnification (using Wolfe 8x20 field microscopes with lens extenders, internal scales graduated in 0.05 mm, and negligible-light-loss non-inverting optics), we are aware of the potential uses of digital technology in type-recurrence analysis. A recent study at Case Western Reserve, for instance, used pattern recognition technology to increase the "yield" of firmly identified individual types using digital Folio images. Perhaps someone with expertise in this area might conduct a similar study using the digitized Hamlet photofacsimiles in the MIT archive. [Back]