WEBVTT 1 00:00:05.360 --> 00:00:08.439 There are forty two books on shelf nineteen B At least, 2 00:00:08.759 --> 00:00:12.960 that is what the catalog says. Yesterday afternoon, after a 3 00:00:13.039 --> 00:00:16.719 lengthy midweek inventory, I enter the number myself on both 4 00:00:16.760 --> 00:00:20.320 my spreadsheet and that little peach posted I keep tucked 5 00:00:20.320 --> 00:00:23.239 at the edge of the metal shelving just above eye level. 6 00:00:23.960 --> 00:00:27.000 The count was neat ordinary, as it should have been 7 00:00:28.079 --> 00:00:31.559 this morning. However, the catalog says forty three, while my 8 00:00:31.600 --> 00:00:35.560 own record, married to the same quiet precision I always apply, 9 00:00:36.119 --> 00:00:39.399 still insists on forty two. When I finally make it 10 00:00:39.479 --> 00:00:43.240 down to the lower archive clipboard in hand, the morning 11 00:00:43.320 --> 00:00:46.159 lull broken only by the rattle of a cart and 12 00:00:46.200 --> 00:00:49.439 the faintest trickle of humidity tugging at my collar, I 13 00:00:49.520 --> 00:00:54.560 count forty five. Forty five spines present and accounted for. 14 00:00:55.520 --> 00:00:59.039 At first, the difference feels insignificant, a hiccup in the 15 00:00:59.079 --> 00:01:04.159 inflation calculator of institutional records. Maybe Lewis reshelved a pair 16 00:01:04.280 --> 00:01:07.599 late last evening after I'd locked up, or perhaps another 17 00:01:07.680 --> 00:01:11.439 grad student fanned through the wrong tray before the doors 18 00:01:11.519 --> 00:01:15.239 creaked shut at five. We don't have enough foot traffic 19 00:01:15.280 --> 00:01:18.560 in the basement to create real chaos. If you ignore 20 00:01:18.599 --> 00:01:23.120 the occasional genealogy enthusiast or a rare, wild eyed faculty 21 00:01:23.159 --> 00:01:27.000 member seeking out the nineteen thirty four city directory. Most 22 00:01:27.120 --> 00:01:30.640 days it's just me, my small arc of colleagues, and 23 00:01:30.680 --> 00:01:34.079 the books themselves. Still I am not the sort to 24 00:01:34.120 --> 00:01:37.599 shrug off small tremors in the equilibrium. I am, as 25 00:01:37.640 --> 00:01:41.760 my partner jokes archive proud, not the boldest or most 26 00:01:41.799 --> 00:01:44.640 loquacious member of the staff, but the one who knows 27 00:01:44.640 --> 00:01:47.159 where to find a paper clip stamped with the library's 28 00:01:47.200 --> 00:01:50.519 nineteen fifties logo, or the single copy of the eighteen 29 00:01:50.599 --> 00:01:56.959 eighty one almanac wedged incorrectly always beside local government reports. 30 00:01:57.560 --> 00:02:00.920 Disparities tug at me. If the count is wrong, the 31 00:02:00.959 --> 00:02:04.200 filing is wrong, and if the filing is wrong, the 32 00:02:04.280 --> 00:02:06.920 system is off balance in a way that nags at 33 00:02:06.959 --> 00:02:10.280 the edges of my piece. I count the shelf again, 34 00:02:10.599 --> 00:02:13.360 this time trailing my finger across the succession of worn 35 00:02:13.439 --> 00:02:17.800 leather cloth and reenforced cardboard. Forty five. I make a 36 00:02:17.800 --> 00:02:21.479 fresh note, then flip through the catalog's accession record. It 37 00:02:21.520 --> 00:02:25.080 promises forty three No. Forty two as of two entries ago, 38 00:02:25.319 --> 00:02:28.400 with a single correction time stamped at midnight, when the 39 00:02:28.400 --> 00:02:32.840 system supposedly locks itself for back ups. My spreadsheet, paused 40 00:02:32.840 --> 00:02:37.360 at last evening's total remains unsullied. For a moment, I 41 00:02:37.400 --> 00:02:41.280 almost laugh. My post it patient and pastal says forty 42 00:02:41.319 --> 00:02:44.919 two rowan w ds reconfirm t h U R s 43 00:02:45.000 --> 00:02:47.919 A M in my tiny upright script. I set a 44 00:02:47.960 --> 00:02:51.360 new one, a neon blue reminder for tomorrow. Some minor 45 00:02:51.439 --> 00:02:55.680 explanation must be waiting, a misallocated interlibrary loan, a lost 46 00:02:55.759 --> 00:02:59.879 unshelved volume tucked behind the wire mesh. My brain, trained 47 00:03:00.240 --> 00:03:04.840 years of benign anomaly catalogs and diagrams every plausible mistake. 48 00:03:05.840 --> 00:03:09.039 But as the quiet settles back in, disturbed only by 49 00:03:09.080 --> 00:03:12.080 the faint vibration of the HVAC and the powdery scent 50 00:03:12.120 --> 00:03:16.639 of cardboard dust, I feel just beneath my organizational instincts, 51 00:03:16.919 --> 00:03:19.919 the spark of something else, the quickening sense that I 52 00:03:20.000 --> 00:03:22.240 might not be able to square the circle this time. 53 00:03:22.879 --> 00:03:26.680 The books, like gentle tricksters, simply refuse to match any 54 00:03:26.719 --> 00:03:29.360 record I give them, and that is a puzzle I 55 00:03:29.360 --> 00:03:33.840 won't let lie. I decide to start fresh tomorrow, routing 56 00:03:33.919 --> 00:03:37.800 each finding in meticulous observation. I will not be defeated 57 00:03:37.840 --> 00:03:40.879 by a stray book, or, as the shelf now seems 58 00:03:40.919 --> 00:03:45.120 to suggest, by three. For all the reputation our library 59 00:03:45.159 --> 00:03:49.080 holds stately solid, a monument of brick with columns that 60 00:03:49.120 --> 00:03:52.879 look harder than marble, its underbelly is anything but immovable. 61 00:03:53.439 --> 00:03:56.879 The archive basement, built in stages between nineteen eleven and 62 00:03:56.919 --> 00:04:00.719 the late sixties, forms a honeycomb of rolling shelves, shadowed 63 00:04:00.759 --> 00:04:04.960 alcoves and half forgotten corridors. In the dusky early hours, 64 00:04:05.240 --> 00:04:08.360 the echoes of footsteps carry oddly, as if the stacks 65 00:04:08.400 --> 00:04:12.360 themselves are listening, or perhaps quietly conversing behind your back. 66 00:04:13.360 --> 00:04:16.040 I moved up through the ranks here, not quickly, but 67 00:04:16.120 --> 00:04:20.800 with steady devotion, morning checklists, accession logs, and rounds of 68 00:04:20.839 --> 00:04:24.680 temperature and humidity inspections that would bore most to tears. 69 00:04:25.160 --> 00:04:28.079 There are days, I admit, when the only real thrill 70 00:04:28.199 --> 00:04:30.879 is the whispered click of a century old lock as 71 00:04:30.879 --> 00:04:34.000 I retrieve an original type script, or the little shiver 72 00:04:34.120 --> 00:04:37.879 of satisfaction from finding that, despite the stubbornness of time, 73 00:04:38.279 --> 00:04:42.480 the records reconcile for another season. Of course, not every 74 00:04:42.519 --> 00:04:46.279 one approaches the work with such reverence Lewis, my assistant, 75 00:04:46.600 --> 00:04:49.040 is new enough to believe that a digital overhaul will 76 00:04:49.040 --> 00:04:53.120 fix all our woes. He spends hours designing barcoded templates, 77 00:04:53.399 --> 00:04:56.439 often distracted by his phones, gentle buzz, or the clatter 78 00:04:56.480 --> 00:05:00.000 from the reading room upstairs. He's earnest, quick to laugh, 79 00:05:00.480 --> 00:05:03.079 and more than willing to listen, though his patience for 80 00:05:03.160 --> 00:05:09.519 spreadsheet nostalgia my approach wears thin on busy days. Doctor Pennington, 81 00:05:09.800 --> 00:05:13.319 our longtime head librarian, remains a fixture even after his 82 00:05:13.399 --> 00:05:17.920 recent retirement, having worked through every technological upheaval the library 83 00:05:17.920 --> 00:05:21.279 has ever known. His pride in the institution is as 84 00:05:21.319 --> 00:05:24.560 sharp as the creases in his weak old cahkies. He's 85 00:05:24.600 --> 00:05:27.839 a man of habit, ritual coffee at the same battered desk, 86 00:05:28.319 --> 00:05:32.439 afternoon walks round the outside perimeter, and stories that unspool 87 00:05:32.439 --> 00:05:37.160 across decades with only the lightest nudge. Mara, meanwhile, is 88 00:05:37.199 --> 00:05:40.720 a graduate student of archives, splitting her weeks between our 89 00:05:40.759 --> 00:05:45.959 manuscript room and a dissertation on interdisciplinary curation. She's meticulous 90 00:05:45.959 --> 00:05:49.439 in her own way, preferring old school index cards, but 91 00:05:49.560 --> 00:05:52.480 not immune to the allure of a good anomaly. She 92 00:05:52.560 --> 00:05:57.079 once told me somewhat shyly overboxed lentil stew. We pretend 93 00:05:57.160 --> 00:06:00.240 cataloging is about closure, but the gaps are where the 94 00:06:00.319 --> 00:06:04.959 magic is. Routine for all these reasons is my cornerstone. 95 00:06:05.360 --> 00:06:08.879 Each day is guided by its circuit review rare donations, 96 00:06:09.120 --> 00:06:12.519 tidy ledges, and whole punched cards. Check the temperature sensors 97 00:06:12.839 --> 00:06:16.360 our rare volumes dislike heat, and make rounds with security 98 00:06:16.360 --> 00:06:18.680 to log any out of place objects near the locked 99 00:06:18.879 --> 00:06:22.920 rare volumes aisle. There are the silent rituals too, closing 100 00:06:22.920 --> 00:06:26.560 the blackout curtains, logging the faintest traces of salt air. 101 00:06:26.600 --> 00:06:29.839 On days when the outside wind shifts weekly in from 102 00:06:29.879 --> 00:06:34.360 the bay, we have a patchwork of record keeping institutional 103 00:06:34.399 --> 00:06:39.000 relics overlapping three or four eras, periodic purges only ever 104 00:06:39.079 --> 00:06:43.319 half finished, the digital catalog running in parallel with battered 105 00:06:43.399 --> 00:06:48.839 ledges and decimalized log books. Old buildings absorb quirks like sponges. 106 00:06:49.199 --> 00:06:51.680 We who tend them simply learn to walk around the edges. 107 00:06:52.079 --> 00:06:55.360 Catalog clocks sometimes run three or four minutes off, even 108 00:06:55.399 --> 00:06:59.040 after replacements. The occasional whiff of sea water drifts up 109 00:06:59.040 --> 00:07:02.160 the north corridor, though the nearest brine is blocks away. 110 00:07:02.720 --> 00:07:07.160 I've carried these oddities, these innocent little fluxes, like a badge. 111 00:07:07.600 --> 00:07:10.920 Until this week. They only ever felt proof of the longevity. 112 00:07:11.240 --> 00:07:15.199 Not the fragility of the archive, but the shelf. That 113 00:07:15.439 --> 00:07:19.240 shelf keeps failing the most basic test. It doesn't fit 114 00:07:19.319 --> 00:07:22.839 any narrative I've learned to trust. On the third day, 115 00:07:22.879 --> 00:07:25.920 I stage a challenge. I wait for the restocking window, 116 00:07:26.120 --> 00:07:30.279 after Lewis leaves, after the day's researchers depart, and count again, 117 00:07:30.480 --> 00:07:33.839 careful and alone. The number is forty four. As I 118 00:07:33.879 --> 00:07:36.800 move through each volume, cross referencing my tally with the 119 00:07:36.839 --> 00:07:40.279 patch together spreadsheet and a wristwatch, my heart SAgs a little. 120 00:07:40.600 --> 00:07:43.079 It should, by every line of logic, be the same. 121 00:07:43.600 --> 00:07:46.040 Yet when I check the digital catalog, I see the 122 00:07:46.120 --> 00:07:49.199 number forty two blinking up at me, a silent rebuke. 123 00:07:50.279 --> 00:07:53.199 This time the update is time stamped to match my 124 00:07:53.279 --> 00:07:56.560 own physical recount, a detail that puts a twinge of 125 00:07:56.600 --> 00:08:01.360 ournease in my chest. One book catches my attention, an 126 00:08:01.360 --> 00:08:04.839 old navy cloth spine with a penciled twenty seven in 127 00:08:04.879 --> 00:08:08.319 a looping hand. The number doesn't fit our shelf system, 128 00:08:08.560 --> 00:08:11.120 nor does it match any accession number in the handwritten 129 00:08:11.199 --> 00:08:15.160 log from my predecessor or in the catalog. It's simply 130 00:08:15.240 --> 00:08:20.720 additional performing due diligence. I pulled previous years shelf logs 131 00:08:20.759 --> 00:08:24.480 for this section, cross checking each entry. The numbers dance, 132 00:08:24.680 --> 00:08:28.680 never settling thirty nine last spring forty six the year before, 133 00:08:29.000 --> 00:08:32.080 an odd forty three, mixed in with comments about recount 134 00:08:32.159 --> 00:08:36.279 to verify suspicious, not a single inventory lines up with 135 00:08:36.320 --> 00:08:39.639 either the catalog or itself for more than three months. 136 00:08:40.320 --> 00:08:42.960 Lewis called down with a grin and a slightly harried 137 00:08:43.000 --> 00:08:46.600 air shrugs. Maybe someone needed a fote rop. I'd guess 138 00:08:46.639 --> 00:08:49.759 it's just where the interns used to stash returns. When 139 00:08:49.759 --> 00:08:52.799 I point out that every usage is logged, every check 140 00:08:52.840 --> 00:08:57.279 out tracked, especially for rare or restricted stacks, his optimism 141 00:08:57.320 --> 00:08:59.559 holds for just a moment before he lets it flutter 142 00:08:59.600 --> 00:09:03.519 off with other laugh old buildings. Right, who keeps old 143 00:09:03.519 --> 00:09:07.799 logs accurate? Anyway? There's nothing in this answer to satisfy me. 144 00:09:08.240 --> 00:09:10.840 For all his warmth and energy, he's not troubled by 145 00:09:10.879 --> 00:09:14.919 the shelf's stubborn refusal to match expectations. But the logs, 146 00:09:15.279 --> 00:09:19.519 both digital and physical, are clear. No one has checked out, 147 00:09:19.720 --> 00:09:22.799 res shelved, or even physically been assigned to this section. 148 00:09:23.039 --> 00:09:28.000 In the date ranges, the count's most sharply contradict that evening, 149 00:09:28.159 --> 00:09:31.080 With staff gone and the soft overhead lighting humming, I 150 00:09:31.120 --> 00:09:33.799 decide to take the problem into my own hands. I 151 00:09:33.879 --> 00:09:37.320 mark each shelf position with a tiny irrasable dot of pencil, 152 00:09:37.679 --> 00:09:41.120 noting the starting and ending titles by hand. Then, on 153 00:09:41.200 --> 00:09:43.759 my break I sketch a small map of the shelf's 154 00:09:43.840 --> 00:09:47.679 layout where each oversized volume rests. The odd tilt of 155 00:09:47.720 --> 00:09:51.000 a binding here, or a rippled, water stained jacket there, 156 00:09:51.519 --> 00:09:54.159 anything that might serve as a bread crumb for tomorrow's 157 00:09:54.200 --> 00:09:58.080 cartography of the unexplained. I leave locking up a little 158 00:09:58.159 --> 00:10:01.120 later than usual, careful to set the new count in 159 00:10:01.200 --> 00:10:04.279 my spreadsheet and pin my notes to a fresh posted 160 00:10:04.519 --> 00:10:07.960 by my office door. I go home, half hoping to 161 00:10:08.039 --> 00:10:12.200 morrow will reveal a dull, prosaic solution. The only thing 162 00:10:12.320 --> 00:10:15.200 worse than a missing book in my conscience is the 163 00:10:15.240 --> 00:10:19.399 sense that logic itself is out of joint. Next morning, 164 00:10:19.519 --> 00:10:22.840 I arrive earlier, armed with my phone's camera and a 165 00:10:22.879 --> 00:10:26.480 fresh note book. The sun is pale through the basement window, 166 00:10:26.799 --> 00:10:30.960 painting chunky rectangles on the cement floor. I photographed the 167 00:10:31.000 --> 00:10:34.320 shelf from the same angle as yesterday, holding my breath 168 00:10:34.360 --> 00:10:38.919 against the faint, almost sweet smell of decaying glue. My 169 00:10:39.000 --> 00:10:42.879 new method is simple. Take a photograph, log the exact 170 00:10:43.039 --> 00:10:47.080 order of visible titles. Mark barely perceptible pencil dots two 171 00:10:47.159 --> 00:10:50.600 millimeters to the right of each spine. I flip through 172 00:10:50.600 --> 00:10:54.200 the shelf with a diligence bordering on obsession, Counting again 173 00:10:54.639 --> 00:10:59.440 forty five. I decide to catalog each title line by line, 174 00:10:59.519 --> 00:11:02.279 on both paper and digital back up. That's when the 175 00:11:02.279 --> 00:11:07.360 contradictions bloom. Several stand out as clearly mismatched. One slim binding, 176 00:11:07.600 --> 00:11:10.600 which yesterday was titled on the Tides and Men, reads 177 00:11:10.600 --> 00:11:13.480 to day as of the Tides in Men. The difference 178 00:11:13.559 --> 00:11:17.879 is tiny, plausible printer's fatigue or a misremembered label. But 179 00:11:17.960 --> 00:11:21.360 standing with a photo from yesterday, the title in yesterday's 180 00:11:21.360 --> 00:11:26.399 capture is objectively, irrefutably different. I try to put it aside, 181 00:11:26.600 --> 00:11:30.080 hunting for better explanations. Some title seems so worn the 182 00:11:30.159 --> 00:11:33.159 lettering itself could have been misread. I check with Mara, 183 00:11:33.360 --> 00:11:36.000 who stops by the shelf later for her research. She 184 00:11:36.120 --> 00:11:39.240 pauses just above the half empty row. That section was 185 00:11:39.320 --> 00:11:42.960 oddly full last week, she says, pulling a blue bound volume, 186 00:11:43.039 --> 00:11:46.080 with no recall of ever seeing it before. I tell 187 00:11:46.080 --> 00:11:48.759 her I've never seen so many blue bindings in that row. 188 00:11:49.200 --> 00:11:51.960 She looks at me with a crooked eyebrow. You sure 189 00:11:53.080 --> 00:11:56.720 I remember the color? It's what caught my eye. I 190 00:11:56.799 --> 00:11:59.799 cross check with securities logbook, the one kept off the 191 00:11:59.840 --> 00:12:02.879 eye chival corridor, near where I sometimes hide to read 192 00:12:02.960 --> 00:12:06.159 during the long winter lull. A note from last week 193 00:12:06.360 --> 00:12:09.799 in the careful script of Simon a night guard, references 194 00:12:09.799 --> 00:12:13.840 cleaning up a spilled cup by the north stacks. I'm certain, 195 00:12:14.159 --> 00:12:17.519 absolutely certain the accident was to the west, near the 196 00:12:17.559 --> 00:12:21.240 old periodical display. Has my memory slipped out of key? 197 00:12:21.279 --> 00:12:26.320 Too determined? Now I catalog every volume by accession and author. 198 00:12:26.720 --> 00:12:29.480 Two leap out D. R. Halworth, the Elm and the 199 00:12:29.519 --> 00:12:33.879 Hour and E. Petra reigns Ashore. Neither name registers in 200 00:12:33.919 --> 00:12:36.919 any library record I can access, not at our own, 201 00:12:37.240 --> 00:12:39.960 not in my union catalogs for the region, not even 202 00:12:39.960 --> 00:12:43.559 as a footnote in dusty indices of obscure authors. They 203 00:12:43.600 --> 00:12:46.720 are apparently unique to this shelf, and perhaps only to 204 00:12:46.799 --> 00:12:51.600 this moment. Unconvinced, I check my photo log just in case. 205 00:12:52.159 --> 00:12:55.080 The snap shot from yesterday shows the spines in roughly 206 00:12:55.120 --> 00:12:58.240 the same order, but with a twist. An old white 207 00:12:58.320 --> 00:13:02.039 volume previously faded to yellow, is now faintly green in 208 00:13:02.080 --> 00:13:05.759 today's image. One of the previously missing books is present, 209 00:13:06.000 --> 00:13:10.200 peeking out, but the title is barely legible. I begin 210 00:13:10.279 --> 00:13:12.799 to notice a pattern, or at least a swirling of 211 00:13:12.840 --> 00:13:17.559 not quite consistent facts. Searching through a digitized database, I find, 212 00:13:17.600 --> 00:13:21.039 in an unrelated text from the nineteen seventies, an obscure 213 00:13:21.080 --> 00:13:25.360 commentary on regional publishing, a fleeting reference to Halworth, whose 214 00:13:25.399 --> 00:13:30.919 revised ELM proved widely influential in neglected circles. But aside 215 00:13:30.919 --> 00:13:34.519 from this stray mention, nothing the author exists in the 216 00:13:34.559 --> 00:13:39.360 margins only. Forensics at this stage seem to multiply confusion. 217 00:13:39.720 --> 00:13:42.480 When I ask Lewis about shelving a particular read book, 218 00:13:42.679 --> 00:13:45.679 he confirms, Yeah, that was just before the break. I 219 00:13:45.720 --> 00:13:48.919 always hated the way it stuck out. Mara, without knowing 220 00:13:48.919 --> 00:13:52.159 our conversation, repeats her earlier memory, The blue one was 221 00:13:52.200 --> 00:13:57.360 always there. My growing spreadsheet collects odd notations, seealt log 222 00:13:57.679 --> 00:14:02.159 titles shifted verify accession record, as if the records themselves 223 00:14:02.240 --> 00:14:06.679 are bracing for instability. The logic I thought immutable, the 224 00:14:06.759 --> 00:14:10.600 certainty that order must win out, feels as if it's unspooling. 225 00:14:11.159 --> 00:14:14.039 I start a private log, too embarrassed to share my 226 00:14:14.120 --> 00:14:17.519 confusion with those who would think me over fastidious. But 227 00:14:17.639 --> 00:14:22.320 I am contrary to every rational instinct entranced. The shelf 228 00:14:22.320 --> 00:14:25.960 has become a laboratory. I call Mara in for her 229 00:14:25.960 --> 00:14:29.480 take on the phenomenon just before closing. She is open minded, 230 00:14:29.759 --> 00:14:32.679 bright eyed, with the faint exhaustion of leave weak research, 231 00:14:33.000 --> 00:14:36.639 and reads my notations with an appreciative hum. If I 232 00:14:36.639 --> 00:14:39.159 didn't know better, I'd say you were tracking a haunting, 233 00:14:39.399 --> 00:14:43.000 she jokes, but a smile wavers with interest. Maybe it's 234 00:14:43.000 --> 00:14:46.919 a software bug, some automated entry gone mad, she adds. 235 00:14:47.360 --> 00:14:51.679 The tone is only half convincing. Of all the grad students, 236 00:14:52.000 --> 00:14:54.480 she is the most likely to follow a story down 237 00:14:54.559 --> 00:14:58.200 to its tangled roots. Yet each time I attempt to 238 00:14:58.240 --> 00:15:02.799 fix rational, record based or digital, the evidence refuses to 239 00:15:02.879 --> 00:15:07.320 lie still. One finds a lost book, another misplaces a title. 240 00:15:07.759 --> 00:15:11.279 Mara and I, looking at the same shelf, see different patterns. 241 00:15:11.720 --> 00:15:16.200 Each act of observation seems to amplify the strangeness. I 242 00:15:16.240 --> 00:15:18.279 am no longer sure if I am hunting for a 243 00:15:18.279 --> 00:15:22.000 plugged in answer, or for something else entirely, a new 244 00:15:22.039 --> 00:15:24.960 class of knowledge, or perhaps a seam in the warp 245 00:15:25.000 --> 00:15:28.919 and weft of the archive. That night, unable to rest, 246 00:15:29.240 --> 00:15:33.440 I linger late, logging my findings in parallel, determined now 247 00:15:33.480 --> 00:15:37.279 not just to record, but to understand. It is well 248 00:15:37.320 --> 00:15:41.080 after closing when I spot doctor Pennington, in his usual rhythm, 249 00:15:41.440 --> 00:15:45.080 making his way through the stacks, coffee mug gripped in 250 00:15:45.120 --> 00:15:48.879 a way that speaks of decades of habit. He pauses 251 00:15:48.919 --> 00:15:52.159 when I call his name, the ceiling's orange light catching 252 00:15:52.200 --> 00:15:56.000 in his thinning hair. He smiles in that half distracted way, 253 00:15:56.240 --> 00:15:59.320 which means his mind is already married to library matters. 254 00:15:59.840 --> 00:16:03.600 I approached gingerly, aware that even as a retired director, 255 00:16:03.840 --> 00:16:09.039 he carries authority like an invisible satchel. Pennington, I say, 256 00:16:09.279 --> 00:16:12.159 trying to keep my tone light. Do you remember anything 257 00:16:12.200 --> 00:16:15.960 peculiar about the shelf in the lower archive nineteen B 258 00:16:16.440 --> 00:16:21.320 near the holding case? He considers anciently deliberate, brushing his 259 00:16:21.360 --> 00:16:25.120 fingers across his chin. Some books just drift, don't they. 260 00:16:25.519 --> 00:16:29.879 His tone is wistful. I res shelved crownings a dozen times, 261 00:16:30.279 --> 00:16:33.159 always on that left side, near the finger jointed panel. 262 00:16:34.240 --> 00:16:37.480 The name teases at me. I have no crownings listed. 263 00:16:38.039 --> 00:16:41.519 Are you thinking of a different section, I ask? He 264 00:16:41.600 --> 00:16:44.039 waves me off, then leans in when I show him 265 00:16:44.039 --> 00:16:48.399 my collection of impossible authors and their apparent books. Yes, yes, 266 00:16:48.639 --> 00:16:52.120 Halworth odd sort, his history of Fog came and went. 267 00:16:52.519 --> 00:16:54.960 I remember a review in the Quarterly back in oh 268 00:16:55.000 --> 00:16:58.600 maybe seventy three. When I probe for more, he only shrugs, 269 00:16:59.080 --> 00:17:03.159 you should look it up. The gap s in the index. Later, 270 00:17:03.320 --> 00:17:07.599 scrolling through the Quarterly's digitized indexes, I find nothing for Halworth. 271 00:17:08.039 --> 00:17:11.960 No review instead in a back issue, a peculiar footnote, 272 00:17:12.200 --> 00:17:17.400 notable gap in regional monographs, shelf location unknown. The review 273 00:17:17.519 --> 00:17:20.839 is as absent as the supposed book. It is, if 274 00:17:20.880 --> 00:17:23.680 Pennington is to be trusted, a review that ought to 275 00:17:23.759 --> 00:17:27.559 exist for a book that disappears with the catalog. My 276 00:17:27.680 --> 00:17:32.200 intuition flips, sliding from the technical to something stranger. Are 277 00:17:32.240 --> 00:17:36.640 these books, these authors, these shifting records, real only when 278 00:17:36.640 --> 00:17:40.200 some one remembers them. What precisely would it mean for 279 00:17:40.240 --> 00:17:42.160 a book to hover on the edge of being by 280 00:17:42.200 --> 00:17:46.920 no fixed agency but collective institution. This is nothing I 281 00:17:46.960 --> 00:17:51.039 can fix with a pencil or post it. I expand 282 00:17:51.119 --> 00:17:55.119 my research, drawing Mara and Lewis in his informal Investigators. 283 00:17:55.680 --> 00:17:59.480 Lewis more digital by inclination helps me run batch searches 284 00:17:59.480 --> 00:18:03.519 for authors. An accession numbers across every public library consortium 285 00:18:03.559 --> 00:18:08.640 we can access. Little comes up, except occasional, tantalizing ghost entries, 286 00:18:08.920 --> 00:18:15.680 temporary holding, citation only, and in one case unverified. Mara's approach, 287 00:18:16.000 --> 00:18:19.599 all networked expertise and doggyed guides leads her to a 288 00:18:19.680 --> 00:18:23.720 rumor that another university once kept a phantom shelf, now 289 00:18:23.759 --> 00:18:28.680 long since disbanded. We build a shared file every instance 290 00:18:28.680 --> 00:18:31.440 of a book or author that refuses to normalize in 291 00:18:31.440 --> 00:18:37.920 the records, every personal recollection that diverges the patterns resist consensus. 292 00:18:38.519 --> 00:18:42.440 The culminating contradiction comes by way of interlibrary loan. I 293 00:18:42.519 --> 00:18:46.279 request a scan of Halworth's history of fog testing Fate. 294 00:18:46.839 --> 00:18:50.640 Three lending libraries claim, in rapid succession to have temporary 295 00:18:50.680 --> 00:18:54.680 records of receiving it, yet no shipping documentation supports any 296 00:18:54.720 --> 00:18:59.720 actual transfer. When I press, one librarian admits almost sheepishly. 297 00:19:00.119 --> 00:19:02.799 It was always a ghost in our files, one of 298 00:19:02.799 --> 00:19:08.079 those books nobody remembers actually holding. Our efforts ripple outward. 299 00:19:08.200 --> 00:19:11.359 I compose a detailed query for a deep background academic 300 00:19:11.440 --> 00:19:15.960 mailing list, outlining the vanishing and reappearing books, the fluctuations 301 00:19:16.000 --> 00:19:19.440 in the shelf, the ghostly authors that exist more convincingly 302 00:19:19.480 --> 00:19:23.240 in memory than on paper. The few responses that trickle 303 00:19:23.279 --> 00:19:27.480 back are uncertain, mostly vague musings, hints that others too 304 00:19:27.880 --> 00:19:31.039 have encountered similar legends, but no one owns up to 305 00:19:31.079 --> 00:19:34.680 having handled one of these entities. If I began with 306 00:19:34.759 --> 00:19:37.720 an obsession to correct a simple error, now I stand 307 00:19:37.720 --> 00:19:40.359 at the edge of a much larger enigma, one whose 308 00:19:40.400 --> 00:19:43.720 boundaries refuse to come into focus. The shelf is not 309 00:19:43.759 --> 00:19:47.799 merely a depot for lost or neglected volumes. Instead, it 310 00:19:47.839 --> 00:19:50.559 appears to be a knot, an impossible point in the 311 00:19:50.599 --> 00:19:54.359 library where the fabric of knowledge itself pushes back. Books 312 00:19:54.400 --> 00:19:59.119 appear not as errors, but as genuinely untraceable phenomena, objects 313 00:19:59.119 --> 00:20:03.039 lodged in a que emulated memory. Perhaps this shelf is 314 00:20:03.119 --> 00:20:06.359 not broken, but singular I can no longer call it 315 00:20:06.400 --> 00:20:09.880 a mistake. Hovering in the blue glow of the stacks. 316 00:20:10.119 --> 00:20:13.519 I begin drafting my next move, realizing that the answers 317 00:20:13.559 --> 00:20:17.440 I seek will demand more than steely record keeping. I 318 00:20:17.519 --> 00:20:21.160 feel fiercely now that our institutional memory has a will 319 00:20:21.599 --> 00:20:25.519 or a whim of its own. Tomorrow I will abandon 320 00:20:25.559 --> 00:20:28.880 the audit and begin a much broader search. Determine to 321 00:20:28.920 --> 00:20:32.359 treat the shelf not as a localized anomaly, but as 322 00:20:32.400 --> 00:20:37.119 the entry point to something altogether stranger. The possibility that this, not, 323 00:20:37.480 --> 00:20:41.000 however small, is a fixed part of the library's reality. 324 00:20:42.039 --> 00:20:45.160 It's already morning before my body finally yields to sleep. 325 00:20:45.720 --> 00:20:48.359 Somewhere in those drifting hours between the hospital light of 326 00:20:48.359 --> 00:20:51.119 the basement and the grainy hush of my living room, 327 00:20:51.319 --> 00:20:55.400 I dream pages flipping of their own accord, shelves stretching 328 00:20:55.440 --> 00:20:58.920 backward into blue shadow, every book and title poised on 329 00:20:58.960 --> 00:21:04.160 the edge of being. My phone alarm shatters the liminal bleary, 330 00:21:04.200 --> 00:21:07.720 But determined, I collect a fresh note book, convince the 331 00:21:07.759 --> 00:21:10.480 old coffee machine to work, and walk back to the 332 00:21:10.519 --> 00:21:13.559 library with the steady sense that, for the first time 333 00:21:13.599 --> 00:21:16.960 in years, I am not working to explain away the mystery, 334 00:21:17.240 --> 00:21:21.079 but to understand its shape. By nine, I have cornered 335 00:21:21.079 --> 00:21:24.839 both Mara and Lewis. Mara, early as always, is sorting 336 00:21:24.880 --> 00:21:27.920 loose folios at a high table near the window. Lewis 337 00:21:27.960 --> 00:21:31.720 is slouched over his monitor, headset crooked, his fingers feeding 338 00:21:31.720 --> 00:21:35.839 the hungry database another batch of barcodes. Neither look surprised 339 00:21:35.880 --> 00:21:38.839 when I ask if we can experiment. There's something about 340 00:21:38.839 --> 00:21:42.039 the shelf that's drawn them both not as skeptics, but 341 00:21:42.079 --> 00:21:45.319 as participants, eager, if a little sheepish, to see how 342 00:21:45.359 --> 00:21:49.359 far beneath the surface this oddity might run. We agree 343 00:21:49.359 --> 00:21:52.920 on simple procedures, hard data, at least as much as 344 00:21:52.920 --> 00:21:56.079 the rules of sanity will allow. All three of us 345 00:21:56.119 --> 00:21:59.720 will count and record Shelf nineteen B alone, and then 346 00:21:59.759 --> 00:22:04.240 to get together, trying to cross reference titles, authors, binding, colour, 347 00:22:04.480 --> 00:22:09.000 and fault lines in memory, the rules into one by one, 348 00:22:09.279 --> 00:22:13.160 separated by ten minutes. No phones or crossedork notes prepared 349 00:22:13.160 --> 00:22:16.160 in isolation. The air between us is as thick as 350 00:22:16.200 --> 00:22:19.519 the silence before a card trick. I go first, forty 351 00:22:19.519 --> 00:22:22.720 five books as before d R. Halworth, the elm and 352 00:22:22.759 --> 00:22:26.079 the hour present in the fourth slot. Rains are shore 353 00:22:26.359 --> 00:22:30.240 blue binding eighth from the left, one red spine, two yellow, 354 00:22:30.440 --> 00:22:34.440 three worn with the absence of colour. When my round ends, 355 00:22:34.480 --> 00:22:37.759 I nod to Mara. She is quick and careful, her 356 00:22:37.799 --> 00:22:42.599 handwriting angled and sharp, flickers through each title, each accession tag, 357 00:22:43.160 --> 00:22:45.839 her notes I'd bet will be more precise than mine. 358 00:22:46.920 --> 00:22:51.759 Lewis last takes the longest, brow knotted, tapping each volume lightly. 359 00:22:52.359 --> 00:22:55.799 When he rejoins us, his face is unreadable, the look 360 00:22:55.839 --> 00:22:59.039 not of confusion, but of someone trying to stitch together 361 00:22:59.200 --> 00:23:03.279 three versions of reality. In the staff lounge, we cross 362 00:23:03.319 --> 00:23:06.480 reference our notes. It doesn't take long for the discrepancies 363 00:23:06.519 --> 00:23:09.839 to pile up. Mara lists forty three. She insists that 364 00:23:09.960 --> 00:23:13.039 rains Ashore is brown, not blue, and the elm and 365 00:23:13.079 --> 00:23:16.880 the hour appears only as elm and time. Lewis, perhaps 366 00:23:16.880 --> 00:23:20.079 hoping for a less uncanny result, has forty four. His 367 00:23:20.240 --> 00:23:23.079 red spine falls in the wrong place. He finds an 368 00:23:23.119 --> 00:23:26.799 unfamiliar name, e Petro's, in what should be Petra's slot. 369 00:23:27.799 --> 00:23:30.599 We each search our own note books for error, laughter 370 00:23:30.720 --> 00:23:34.400 covering awkwardness that quickly slides into awe. Is it me, 371 00:23:34.839 --> 00:23:38.119 Mara asks, or are the notes themselves starting to argue 372 00:23:38.119 --> 00:23:41.519 with us? Louis says slowly, Maybe the shelf knows whose 373 00:23:41.599 --> 00:23:45.079 counting he tries for a joke, but the line lands 374 00:23:45.119 --> 00:23:49.319 with more reverence than humor. I propose one more experiment. 375 00:23:49.720 --> 00:23:53.079 We'll all enter together, count out, loud riffle each title, 376 00:23:53.480 --> 00:23:57.960 arguing discrepcies in real time. There is something bracing, even giddy, 377 00:23:58.200 --> 00:24:00.680 in standing shoulder to shoulder A mins. It's the dust 378 00:24:00.880 --> 00:24:03.799 and the mild, mineral air, as if preparing for a 379 00:24:03.839 --> 00:24:08.839 precise conjuring. Somehow, The list converges forty four this time, 380 00:24:09.039 --> 00:24:11.960 but no consensus on half the titles, and the blue 381 00:24:12.000 --> 00:24:15.920 binding seems to pale as we debate its pedigree. The 382 00:24:15.960 --> 00:24:19.440 book in e Petra's slot becomes to Mara, a translation 383 00:24:19.960 --> 00:24:23.480 to Lewis a misfiled city register to me, the same 384 00:24:23.640 --> 00:24:27.839 enigmatic volume as before. I reach out, tap the spine, 385 00:24:28.279 --> 00:24:30.960 and for an instant, the air fizzles, as if our 386 00:24:31.000 --> 00:24:34.880 collective observation is creating or erasing the book as we speak. 387 00:24:35.880 --> 00:24:40.279 The moment passes, the shelf resettles into normalcy or something adjacent, 388 00:24:40.599 --> 00:24:42.839 and the three of us stand shaken in the silence, 389 00:24:43.119 --> 00:24:46.960 trading glances that hint at belief, disbelief, and something deeper, 390 00:24:47.359 --> 00:24:51.319 amusement laced with humility. Maybe libraries just accrue their own 391 00:24:51.400 --> 00:24:55.599 quantum rules. After a hundred years. Lewis murmurs. Mara nods 392 00:24:55.640 --> 00:24:59.880 half smiling. Wouldn't that make us all unreliable narrators? Living 393 00:25:00.079 --> 00:25:04.160 indexes Later, I reach out to doctor Pennington, who is 394 00:25:04.200 --> 00:25:07.240 happy to reminisce, but less eager to speculate. I ask 395 00:25:07.279 --> 00:25:10.000 about phantom books. If in his decades here he's known 396 00:25:10.079 --> 00:25:13.160 volumes that seem only half real. He SIPs his coffee 397 00:25:13.160 --> 00:25:17.160 and shrugs. Some books their presence lingers longer than their covers. 398 00:25:17.480 --> 00:25:19.839 I wouldn't be surprised if the building itself keeps a 399 00:25:19.839 --> 00:25:23.559 few around for company. He admits to ras, shelving phantom 400 00:25:23.720 --> 00:25:27.079 monographs too many times to count, always chalking it up 401 00:25:27.160 --> 00:25:30.440