卷:卷三《阳关以西》|主角:Marco·阿里木·阿不都|古人:僧一行|地点:吐鲁番·第二第三第九座晾房·阿不都家院子|典:《旧唐书·天文志》一行“乃诏太史监南宫说等于河南平地测日晷及极星,夏至日中立八尺表”,遣使十三处同测;又《大衍历》“凡日晷……自阳城至朗州,其差三百五十一里八十步而极差一度”|主导感官:温(三座房里三支温度计,同一个时辰,三个数)|合券进度:未结一笔|金融-国学对应:Measure in Many Places on One Day(一处测出来的,只在那一处准)↔ 十三处测影
段一·现身
七月十一日,上午七点。
吐鲁番。
跟去年一样,出机场那一下人站住了。
四十一度。
去年是四十三。
阿里木来接的。
还是那辆皮卡。
车上他先说了一句:
“我爸在家。”
马可问:
“身体怎么样?”
阿里木说:
“能走。”
这三个字答得跟去年不一样。
去年阿里木说他爸的时候,说的是「我爸让我给你」「我爸说要出」。
今年他说的是「能走」。
马可没有接着问。
车往东北走,路两边先是戈壁,然后忽然绿了。
开了四十分钟,土路,晾房。
这一次马可数到第九座就停了。
因为第九座门口挂着一块木牌。
是新的,木头还是白的。
上头用红漆写了三个字。
马可下车走过去看。
写的是:
第九座
段二·古文镜像
“开元十二年,诏太史监南宫说等,于河南平地测日晷及极星。夏至日中,立八尺之表。”
“凡此十三处,皆一日同测。”
“自阳城至朗州,其差三百五十一里八十步,而极差一度。”
English Mirror: In the twelfth year of Kaiyuan the astronomer Nangong Yue and others were ordered to measure the sun's shadow and the pole star on the level ground of Henan. At the summer solstice noon, an eight-foot gnomon was set up. — All thirteen stations measured on one and the same day. — From Yangcheng to Langzhou the distance differs by three hundred and fifty-one li and eighty paces for one degree of polar altitude.
那块牌子是六月做的。
阿里木说的。
他说:
“六月里做了三块。”
马可问:
“第二第三也有?”
阿里木说:
“有。”
马可往回走了几十米,看见了。
第二座和第三座门口,也各挂着一块。
一样的木头,一样的红漆。
马可问:
“别的房子有吗?”
阿里木说:
“没有。”
他说:
“别的房子门框上头有白漆写的号。”
马可点了点头。
去年他就是看着白漆那些号数过去的。
他问:
“为什么这三座要挂牌?”
阿里木说了一句,说得很平常:
“白漆刷一遍就没了。”
那天上午,马可进了第九座。
七月,房子是空的,架子上什么也没有。
跟去年七月一样。
只有一样不一样。
墙上钉了一个东西。
是一支温度计,很普通的那种,玻璃管。
钉在东面墙上,离地一米五。
马可看了一眼。
三十一度。
外头四十一。
他出来问阿里木:
“这个也是六月装的?”
阿里木说:
“四月。”
他说:
“三座房都有。”
马可问:
“谁让装的?”
阿里木说:
“我自己。”
马可问:
“为什么想到装这个?”
阿里木说:
“去年那三天。”
他说:
“第九座比前两座早三天下架。”
他说:
“我爸说是风。”
他停了一下。
他说:
“我想看看是不是。”
马可问:
“为什么不量风?”
阿里木说:
“量风的表贵。”
他说得很实在。
他说:
“温度表一支十二块。”
马可算了一下。
三支,三十六块。
他问:
“这三十六块是您出的?”
阿里木说:
“是。”
马可问:
“为什么不跟我说?”
阿里木想了一下。
他说:
“三十六块。”
这三个字说完,两个人都没有再说。
马可在本子上写了一行,写的时候他有点不舒服:
「他花三十六块,量了一百天。」
他停了一下,写完:
「我这一年,花了一千六重印袋子。」
又写了一行:
「那一千六买的是七行字好看。」
再写一行:
「这三十六买的是一张表。」
写完这四行,马可从钱包里拿了钱。
他数了一下,给了阿里木一百。
阿里木说:
“三十六。”
马可说:
“一百。”
阿里木说:
“三支表三十六。”
马可说:
“表三十六,量了一百天,一天六毛四。”
阿里木停了一下。
他把那一百块接过去了。
接过去以后他说了一句:
“那我明年还量。”
马可说:
“明年不止三座。”
阿里木抬起头。
马可说:
“我想在东头再放三支。”
阿里木问:
“哪三座?”
马可说:
“十八,二十八,三十七。”
阿里木想了一下。
他说:
“十八到三十七,走一趟一里多。”
马可说:
“一天一趟。”
阿里木说:
“一天一趟,二十分钟。”
他没有说行还是不行。
他问的是:
“你要东头那三个数做什么?”
马可说:
“我要知道第九座是不是特别。”
阿里木说:
“那三座又不给你。”
马可说:
“不给我。”
他说:
“可是不知道那二十八座是什么样,我就不知道我那三座是什么样。”
阿里木没有立刻答。
他把那一百块折起来,放进口袋。
过了一会儿,他说了一句:
“这个我跟我爸说。”
那天晚上吃饭的时候,阿里木说了。
他说得很短,就一句:马可想在东头再放三支表。
阿不都在喝汤,没有抬头。
他喝完那一口,问了一句:
“哪三座?”
阿里木说:
“十八、二十八、三十七。”
阿不都想了一下。
他说:
“十八那一座,二〇一一年盖的。”
他说:
“二十八,二〇一七。”
他说:
“三十七,二〇二三。”
马可停住了。
因为这三个年份,阿不都答得跟报自家孩子的生日一样快。
他问:
“阿不都大哥,您每一座都记得是哪一年盖的?”
阿不都说:
“记得。”
马可问:
“一百四十七座?”
阿不都说:
“自己盖的那一百二十七座,记得。”
他说:
“租的那二十座,不记得。”
马可在那张桌子前面坐着,一时没有说话。
因为下午他刚在本子上写过一句「他走不完的那些,他只知道数」。
这一句现在要改。
他走不完,可是他记得每一座是哪一年盖的。
马可问了一句:
“大哥,那您知道十八座跟第九座差多少吗?”
阿不都说:
“不知道。”
他答得很干脆。
他说:
“盖的年份我记得。”
他说:
“里头什么样,我说不上来。”
他把碗放下。
他说了一句,马可当天晚上在本子上抄了三遍:
“盖的时候我在。”
他停了一下。
“盖完以后我没有天天进去。”
马可在本子上,把下午那一行改了。
他划掉的是:「他走不完的那些,他只知道数。」
他重写的是:
「盖的时候他在,所以他记得年份。」
下面一行:
「盖完他不进去,所以他不知道里头。」
再下面一行:
「记得的是他干过的那一段。」
这一行写完,他停了很久。
因为这一句去年十一月裴矩在张掖说过。
一个人只认他自己走过的那一段。
当时那一句是用来问客商的。
今天这一句用在一个盖了一百二十七座房的人身上。
同一句话。
马可在最后写了一行:
「他走过的那一段,是盖房子那几天。」
他写完了:
「不是那房子后来的三十年。」
段三·事件主体
那天上午,阿里木把去年那本拿出来了。
二〇二八年那一本,从七月二十号记到十一月二十号。
十二条回音,加上每天的记录。
他翻到九月里那几页。
他说:
“去年下架的时候,我记了一件事。”
马可说:
“第九座比前两座干得快,三天。”
阿里木说:
“对。”
他说:
“那时候我不知道为什么。”
他说:
“我爸说东头风大。”
马可说:
“风大干得快。”
阿里木说:
“我今年量了。”
他从口袋里掏出一张纸。
是折了四折的一张A4,上面是手抄的表。
三列。
第二座,第三座,第九座。
下面是日子,四月一号到七月十号。
每一天三个数。
那张表,马可看了很久。
他先看平均。
四月:第二座十九点一,第三座十九点三,第九座十八点四。
五月:二十四点七,二十四点九,二十三点八。
六月:三十点二,三十点四,二十九点一。
七月前十天:三十四点八,三十五点零,三十三点六。
第九座每一个月都低。
低多少?
四月低零点八,五月低一点一,六月低一点二,七月低一点三。
马可把这四个数写下来。
零点八,一点一,一点二,一点三。
天越热,差得越多。
他抬起头,问了一句:
“阿里木,这个是风吹出来的?”
阿里木说:
“我不知道。”
他说:
“我只知道低。”
僧一行是那天下午出现的。
中午吃完饭,阿不都睡了。
阿里木去地里。
马可一个人在院子里那张矮桌前面,抄那张表。
葡萄架的影子在桌上,一格一格。
一个人坐在对面。
三十来岁,很瘦。
穿的是僧衣,可是外头罩了一件很旧的圆领袍。
他手里拿着一根尺,木头的,很长。
他说的第一句是:
“三个数,一天量一次。”
马可说:
“是。”
那个人说:
“一行。”
那天下午,一行说了他那件事。
他说:
“开元十二年,我要改历。”
马可问:
“改什么?”
一行说:
“旧历上有一句话。”
他说:
“说是南北相隔一千里,日影差一寸。”
马可问:
“这句话不对?”
一行说:
“这句话用了七百年。”
他说:
“从汉朝就这么说。”
他把那根尺放在桌上。
他说:
“七百年,没有人量过。”
马可停了一下。
他说:
“七百年没有人量?”
一行说:
“量过。”
他说:
“在洛阳量过,在长安量过。”
他抬起头。
他说了一句,马可当天在本子上抄了三遍:
“都是一处一处量的。”
他说:
“这一处量了,那一处也量了,可是不是同一天。”
马可问:
“不同一天有什么关系?”
一行说:
“日影一天一个样。”
他说:
“你今天在洛阳量一个数,明年在长安量一个数。”
他说:
“这两个数一减,”
他把手一摊。
他说:
“减出来的是什么,谁也不知道。”
那天下午,一行说了他怎么做的。
他说:
“我请了十三处。”
马可问:
“十三个地方?”
一行说:
“十三个。”
他说:
“南到交州,北到铁勒。”
他说:
“每一处立一根八尺的表。”
他把那根木尺竖起来。
他说:
“夏至那一天正午,同时量。”
马可问:
“同时?”
一行说:
“同一天。”
他说:
“这一条最要紧。”
他抬起头。
他说了一句,马可当天在本子上抄了四遍:
“十三个数,要是十三天量的,那就是十三个数。”
他停了一下。
“同一天量的,那才是一条线。”
马可在那张桌子前面坐着,把这句想了一遍。
他说:
“先生,量出来的跟旧书上那一句差多少?”
一行说:
“差得远。”
他说:
“旧书说一千里差一寸。”
他说:
“我们量出来,三百五十一里差一度。”
他说:
“换算过来,一千里差不到三分。”
马可算了一下。
一寸是十分。
三分对一寸。
差三倍多。
他抬起头。
他说:
“先生,那七百年里,用这一句算过多少东西?”
一行说:
“很多。”
他说得很平常。
他说:
“算地,算历,算日食。”
他停了一下。
他说:
“都是错的。”
那天下午,马可问了他真正想问的。
他把阿里木那张表推过去。
他说:
“先生,这三列,是同一天量的。”
一行看了一眼。
他说:
“同一个院子。”
马可说:
“是。”
一行说:
“三座房挨着?”
马可说:
“两座挨着,一座在东头。”
一行说:
“隔多远?”
马可说:
“第二第三挨着,第九在那一排的尽头,隔两百多米。”
一行把那张纸看了一会儿。
他说:
“两百米,差一度三。”
他抬起头。
他说了一句,马可当天下午就给阿里木打了电话:
“你这三个数是对的。”
他说:
“可是你这一排房,有多少座?”
马可当时没有立刻答。
因为他以为自己知道。
他说:
“先生,这一排三十几座吧。”
一行说:
“你去数。”
马可说:
“去年我数过。数到三十七我放弃了。”
一行说:
“你数的是三十七。”
马可说:
“是。”
一行说:
“那第九座在第几个?”
这一句问出来,马可停住了。
因为他从来没有把这两件事放在一起。
「第九座」这三个字,他叫了一年。
「东头那一座」,他也说了一年。
他一直以为这两个说的是同一个位置。
一行说:
“你今天下午去数一遍。”
他把那根木尺收起来。
他说:
“我那十三处,最要紧的不是同一天。”
马可抬起头。
一行说:
“最要紧的是我知道每一处在哪儿。”
他说:
“从阳城到朗州多少里,我都量过。”
他说了一句,马可当天在本子上抄了三遍:
“同一天量了十三个数,可是不知道十三处在哪儿,”
他停了一下。
“那还是十三个数。”
马可把这句抄下来。
抄完他问了一句:
“先生,那十三处,量完以后呢?”
一行说:
“量完以后,我要写一句话。”
马可问:
“写什么?”
一行说:
“写这一句是在哪儿量的,什么时候量的,谁量的。”
他说:
“这三样缺一样,那个数就没有用。”
马可问:
“为什么谁量的也要写?”
一行说:
“因为量错了要找得着人。”
他说得很平常。
他说:
“十三处里,有两处的数不对。”
马可问:
“怎么发现的?”
一行说:
“排起来不成线。”
他用手在桌上画了一条线。
他说:
“十一个点在一条线上,两个点不在。”
马可问:
“那两处是怎么回事?”
一行说:
“一处的表立歪了。”
他说:
“一处的人报错了时辰。”
他抬起头。
他说:
“这两处我都问到了人。”
他说了一句,马可当天在本子上抄了两遍:
“要是不写谁量的,我只能把那两个数扔了。”
他说:
“扔了就是十一处。”
他停了一下。
“问着了人,还是十三处。”
段四·钩子结尾
那天傍晚,阿里木从地里回来。
马可在院门口等他。
他问的第一句是:
“阿里木,你们那一排一共多少座?”
阿里木说:
“三十七座。”
马可问:
“从西头到东头多长?”
阿里木想了一下。
他说:
“一里多。”
马可问:
“第二座在哪儿?”
阿里木说:
“从西头数第二个。”
马可问:
“第九座呢?”
阿里木说:
“第九个。”
马可停住了。
他一直以为第九座在那一排的尽头。
因为去年阿不都说的是「一座在东头」。
他问:
“第九座不是东头?”
阿里木说:
“不是。”
他说:
“东头是三十七座。”
那天傍晚,马可站在那条土路上,把这件事重新算了一遍。
三十七座房,一里多,从西到东。
第二座在西边第二个。
第九座在西边第九个。
中间隔七座,两百多米。
东头还有二十八座。
这二十八座,没有一支温度计。
马可问了一句:
“阿里木,最东头那一座,你量过吗?”
阿里木说:
“没有。”
马可问:
“为什么没量?”
阿里木说:
“那三座是给你的。”
他说得很自然。
他说:
“别的房子跟你没关系。”
马可在那条土路上站着,看着那一排房。
三十七座,一里多。
他这两年做的所有事,都在前九座里。
后头二十八座,他一次也没有进去过。
他在本子上写了一行:
「我以为我在做一个对照试验。」
他停了一下,写完:
「两座挨着,一座在东头。」
又写了一行:
「其实那一座不在东头。」
再写一行:
「它在中间。」
这一行写完,马可又在下面写了一句:
「而且我不知道中间跟东头差多少。」
他停了一下,写完:
「因为东头那二十八座,一个数也没有。」
又写一行:
「我这两年做的所有判断,都是九分之一。」
这一行他写完,觉得说得太重,划掉了。
他重写了一句:
「不是错。是范围比我以为的小。」
再写一行:
「而且这个范围,是别人给我划的,我一直以为是我自己划的。」
这一行写完,他抬起头。
因为这意味着阿不都去年七月那句话不是随口说的。
「两座挨着的出的东西一样,东头那一座跟这两座不一样。」
要是第九座真在东头,那这句是一个设计。
第九座在第九个,那这句是什么?
马可回到院子里,阿不都在葡萄架下面坐着。
这是这一天马可第一次见他。
他比去年瘦了一点。
白衬衫,袖子还是挽到手肘。
手上还是那些老茧,指甲还是修得很干净。
马可坐下,问了那一句:
“阿不都大哥,去年您说第九座在东头。”
阿不都说:
“是。”
马可说:
“它在第九个。”
阿不都点了点头。
他说:
“对。”
马可问:
“那为什么说东头?”
阿不都说了一句,马可当时以为自己听错了:
“我记的时候,那一排只有九座。”
马可停住了。
他问:
“哪一年只有九座?”
阿不都想了一下。
他说:
“九六年。”
他说:
“九六年那一排九座,第九座就在东头。”
他说:
“后来一年一年往东盖。”
他抬起头。
他说:
“盖到三十七。”
院子里那棵桑树上有鸟。
马可没有说话。
过了一会儿,阿不都自己说了一句。
他说得很平常,就跟说天气一样:
“我脑子里那一排,还是九座。”
马可在那个院子里,很久没有说话。
过了一会儿,他问了一句:
“阿不都大哥,那您平时说房子的时候,说的是哪一排?”
阿不都说:
“说数的时候,我说一百四十七座。”
他说:
“说位置的时候,我说九座那一排。”
马可问:
“这两个不打架吗?”
阿不都说:
“不打架。”
他说得很平常。
他说:
“一百四十七是数出来的。”
他说:
“九座是走出来的。”
马可停了一下。
阿不都说:
“九六年那一排,我一天走三趟。”
他说:
“从西头走到东头,二百来米。”
他说:
“哪一座今天开了半边洞,哪一座昨天没关严,我走一趟就知道。”
他抬起头。
他说了一句,马可当天在本子上抄了两遍:
“现在三十七座,一里多。”
他说:
“我走不完了。”
院子里那棵桑树上,鸟还在叫。
阿不都说:
“走不完的那些,我只知道数。”
他说:
“走得完的那些,我知道它们是什么样。”
他停了一下。
他说了后半句,马可后来在飞机上想了一路:
“你要的那三座,在我走得完的那一段里。”
马可在那张矮桌前面坐着,把这一句想了很久。
去年七月,他要三座房单记。
阿不都当场就说了「两座挨着,一座在东头」。
当时马可觉得这是一个设计。
今天他知道,这也是一个范围。
阿不都给他的,是他自己还走得完的那九座。
不是因为那九座最好。
是因为那九座,他闭着眼睛也知道是什么样。
要是给东头那二十八座里的三座,
阿不都自己也说不出那三座跟别的有什么不一样。
马可在本子上写了一行:
「他给我的不是最好的三座。」
他停了一下,写完:
「是他还答得出话的三座。」
又写了一行:
「一个人能给你的,只有他答得出的那一部分。」
这一行写完,他忽然想起去年十一月周立行说的那句。
我在吐鲁番待了四天,什么也没问出来。
当时马可以为那是周立行不会问。
今天他想到另外一种可能。
周立行问的,可能正好是那二十八座。
问那二十八座,谁也答不出。
马可把这个写下来了。
写完他在旁边加了一句:
「这个我没有证据。」
欲知后事如何,请看下回,《一里多》
附录一·Dongpo's Note
Out of the airport at seven and I stopped again, as last year. Forty-one today; forty-three then. Alim in the same pickup, and the first thing he said was my father's at home. How is he? He can walk. Which is not how he spoke about him last year — last year it was my father says to give you this, my father says we should pay it. This year it is he can walk. I let it go.
Forty minutes, dirt track, drying houses. I counted to nine and stopped, because there is a wooden board hung by the door of house nine, new, the wood still pale, three characters in red paint: house nine.
Made in June, and two more like it on houses two and three. Nothing on any of the others — they carry the number in white paint above the door frame, which is what I counted along last year. Why boards on these three? Entirely level: white paint goes with one coat of whitewash.
Inside house nine, empty as it is every July, one thing is new: a plain glass thermometer nailed to the east wall at a metre and a half. Thirty-one inside; forty-one out. Also June? April. All three houses? All three. Who told you to? Nobody. I did it.
Then he produced last year's book and turned to September. When we took down last year I wrote something down. House nine dried three days faster than the other two. Yes. I didn't know why then. My father said more wind at the east end. And: this year I measured it.
A sheet of A4 folded in four, hand-ruled, three columns — houses two, three and nine — and a row for every day from 1 April to 10 July, three readings each.
Monthly means. April: 19.1 / 19.3 / 18.4. May: 24.7 / 24.9 / 23.8. June: 30.2 / 30.4 / 29.1. First ten days of July: 34.8 / 35.0 / 33.6.
House nine is lower every month, and the gap widens: 0.8, 1.1, 1.2, 1.3. The hotter it gets, the bigger the difference. Is that the wind? I don't know. I only know it's lower.
Yixing was opposite me at the low table after lunch while the household slept — thirty-odd, very thin, monk's robe with an old round-collared coat over it, holding a long wooden rule. Three readings, once a day. Yes. Yixing.
The twelfth year of Kaiyuan, and a calendar to revise. What needed revising? There was a sentence in the old calendar. It said that a thousand li north or south changes the sun's shadow by one cun. Wrong? It had been in use for seven hundred years. Since the Han. And, putting the rule down: in seven hundred years nobody had measured it.
Nobody? They measured. At Luoyang, at Chang'an. And then the line: they measured one place at a time. And: this place measured, that place measured — but not on the same day. What difference does that make? The shadow is different every day. You take a reading at Luoyang today and a reading at Chang'an next year, and you subtract them — hands apart — and nobody knows what the difference is.
So: thirteen stations. From Jiaozhou in the south to the Tiele in the north. An eight-foot gnomon at each. At noon on the solstice, all at once. Simultaneously? On the same day. That is the part that matters. And: thirteen readings taken on thirteen days are thirteen readings. Taken on one day, they are a line.
How far out was the old sentence? A long way. The books said a thousand li to a cun. We measured three hundred and fifty-one li to one degree — which comes to under three fen per thousand li, against one cun, which is ten. More than three times. And in seven hundred years, how much had been computed from it? A great deal. Land, calendars, eclipses. A pause. All of it wrong.
I pushed Alim's sheet across. These three columns are from the same day. He looked. Same courtyard. Yes. Are the three houses next to each other? Two adjacent, one at the east end. How far apart? Two and three adjoin; nine is at the end of the row, two hundred-odd metres away. He studied it. Two hundred metres, one point three degrees. And then: your three numbers are right. And: but how many houses are there in that row? — M.
附录二·Marco's Memo
Memo to file. Turpan, 11 July 2029.
1. Three things Alim has done without being asked. — Wooden name boards on houses two, three and nine, June, red paint. Reason: white paint goes with one coat of whitewash. No other house in the row has one. — A glass thermometer nailed to the east wall of each of the three, April, at 1.5 m. — A daily three-column temperature series, 1 April to 10 July, on one hand-ruled sheet folded in four in his pocket. None of this is on the record sheet I specified. He has extended the instrument on his own account for a second consecutive year (ch.164's eighth column was Xiao Lei doing the same thing in Chengdu).
2. The series. Monthly means for houses 2 / 3 / 9: April 19.1 / 19.3 / 18.4; May 24.7 / 24.9 / 23.8; June 30.2 / 30.4 / 29.1; July (1–10) 34.8 / 35.0 / 33.6. — House nine is cooler every month, and the deficit widens with the season: 0.8, 1.1, 1.2, 1.3. — Which is consistent with last September's observation that nine dried three days faster, and gives it a mechanism, or at least a correlate. Alim's own position, correctly stated: I don't know. I only know it's lower.
3. Yixing, and the methodological point is the whole chapter. — The received figure — a thousand li changes the shadow by one cun — had stood seven hundred years since the Han, and had been used for land measurement, calendars and eclipse prediction. All of it wrong. — It was not that nobody had measured. They measured one place at a time, on different days. And the shadow changes daily, so differencing a Luoyang reading against a Chang'an reading taken a year apart produces a number of unknown composition. — Thirteen stations, Jiaozhou to the Tiele, eight-foot gnomons, all at noon on the same solstice. — Thirteen readings taken on thirteen days are thirteen readings. Taken on one day, they are a line. — Result: 351 li 80 paces per degree, i.e. under three fen per thousand li against a received ten. Out by more than three times.
4. And his question, which undid what I thought I had. He read the sheet and said the three numbers are right, and then asked how many houses are in the row. — Thirty-seven. Over a li from the west end to the east. — House two is second from the west. House nine is ninth from the west. They are 200-odd metres apart, and twenty-eight houses lie further east, none of them measured. — House nine is not at the east end. It is a quarter of the way along.
5. Which means my control-and-variable design was never what I was told it was. Abdul Rahman, July 2028: two adjacent, one at the east end; the two adjacent ones produce the same thing, the east one does not; more wind at the east end, dries three or four days faster. I recorded that as two controls and one variable and have cited it four times since. It is still a real design — nine differs measurably from two and three — but the axis is not "end of the row." It is "two hundred metres east of two houses that adjoin each other," in a row that continues for another kilometre. Everything I have concluded about the effect of position is bounded by nine of thirty-seven buildings, and I did not know that until this afternoon.
6. The reason, and it is not an error. Asked why he called it the east end when it is ninth of thirty-seven: when I got used to it, the row had nine houses. Which year? '96. In '96 the row was nine and nine was the east end. And then it was extended eastward, year after year, to thirty-seven. And, unprompted, in the tone of a remark about weather: the row in my head is still nine.
7. What to do about it, and I am not going to over-correct. The finding stands as far as it goes: house nine runs 0.8–1.3° cooler than houses two and three, measured daily on the same days by one hand, and dries three days faster. What it does not establish is anything about position along the row, because there is no reading east of the ninth building. Cheapest fix: three more thermometers — say houses 18, 28 and 37 — and the same daily sheet. Cost of a glass thermometer is a few yuan; the cost is Alim's walk, which is a kilometre each way. I will ask, and I will ask for it as work and price it, not as a favour — ch.158, and Abdul Rahman's rule about what is given for nothing.
8. And the thing underneath all of this, which is item 1 of ch.157 arriving with a date on it. In 2027 Abdul Rahman told his youngest son I can't remember any more, and that is why the notebooks restarted. What he cannot remember is not names or prices. It is that the row grew. He is describing, accurately and in good faith, a courtyard that stopped existing in about 1997. Everything I have built here rests on his descriptions. They are true of the place he learned. I have been treating a man's account of his own yard as ground truth for two years, and today one sentence of it turned out to be thirty-two years out of date. The instrument that caught it was three columns of temperature written down by his son.
附录三·English Highlight
"Thirteen readings taken on thirteen days are thirteen readings. Taken on one day, they are a line."
The sentence that needed correcting had been in the calendars since the Han: a thousand li north or south changes the length of the noon shadow by one cun. Seven hundred years, and land had been measured with it, and calendars built on it, and eclipses predicted from it.
It was not that nobody had measured. Readings had been taken at Luoyang and at Chang'an and elsewhere, carefully, by competent people. They were taken one place at a time, and on different days — and the shadow changes daily, so subtracting a reading made in one city from a reading made in another a year later yields a number of unknown composition.
So thirteen stations were established, from Jiaozhou in the far south to the Tiele in the north, each with an eight-foot gnomon, all measuring at noon on the same solstice. The answer came out at three hundred and fifty-one li and eighty paces to one degree of polar altitude — under three fen per thousand li where the books had said ten. Out by more than a factor of three. And everything computed from it in seven centuries: all of it wrong.
In a courtyard in Turpan a thirty-one-year-old had nailed a glass thermometer to the east wall of each of three drying houses in April, unasked, and had been writing down three readings a day ever since on one sheet of paper folded in four in his pocket. Same day, same hand, three places. House nine ran cooler than houses two and three every month, and the gap widened with the heat: eight-tenths of a degree in April, one and three-tenths in July.
Shown the sheet, the man who had organised the thirteen stations said the three numbers were right, and then asked how many houses were in the row.
Thirty-seven. Over a kilometre from one end to the other. House two is second from the west; house nine is ninth from the west; and twenty-eight further buildings run east of it, none of them measured.
The arrangement had been described the previous summer as two adjacent and one at the east end. It is not at the east end. Asked about it, the man who said it did not correct himself and did not apologise.
When I got used to it, the row had nine houses. In 1996. And then it was extended eastward, year after year.
The row in my head is still nine.
附录四·Dongpo's Cheat Sheet
| 七月十一 | 出处 | 现代对应 | 一句话 |
|---|---|---|---|
| 我爸在家。身体怎么样。能走 | 去年他说的是「我爸让我给你」 | 今年是「能走」 | |
| 三块木牌,红漆 | 六月做的 | 别的房子门框上是白漆的号 | 白漆刷一遍就没了 |
| 三支温度计,四月装的 | 谁让装的。我自己 | 一天三个数,四月到七月 | 折了四折在口袋里 |
| 第九座每个月都低 | 0.8/1.1/1.2/1.3 | 天越热差得越多 | 我不知道为什么,我只知道低 |
| 一千里差一寸 | 用了七百年,从汉朝起 | 七百年没有人量过 | 量过,可是一处一处量的 |
| 日影一天一个样 | 今天洛阳一个数,明年长安一个数 | 一减,减出来的是什么谁也不知道 | |
| 十三处,南到交州北到铁勒 | 八尺表,夏至正午 | 同一天 | 十三天量的是十三个数,同一天量的才是一条线 |
| 三百五十一里差一度 | 一千里不到三分,旧书说一寸 | 差三倍多 | 七百年算的地、历、日食,都是错的 |
| 你这三个数是对的 | 可是你这一排房有多少座 | ||
| 三十七座,一里多 | 第九座是西头第九个 | 东头还有二十八座,一支温度计也没有 | 第九座不在东头,在四分之一处 |
| 那三座是给你的 | 别的房子跟你没关系 | ||
| 我记的时候那一排只有九座 | 九六年,那时候第九座就在东头 | 后来一年一年往东盖,盖到三十七 | 我脑子里那一排,还是九座 |
| 他记不住了 | 不是名字和价钱 | 是那一排长过 | 两年的地基,是一个一九九七年就不在了的院子 |
| 抓住这一条的 | 是他儿子那三列温度 |