卷:卷三《阳关以西》|主角:Marco·郭老师·常玉山(消息)|古人:贾思勰(《齐民要术》)|地点:成都·马可的出租屋窗台·郭老师家车库|典:《齐民要术》卷四“作乾蒲萄法:极熟者一一零叠取,刀子切去蔕,勿令汁出……漉出,阴乾便成矣”;又其序“采捃经传,爰及歌谣,询之老成,验之行事,直取会意,不尚浮辞”|主导感官:黏(第十七天,两个葡萄干捏一下不分开)|合券进度:半张|金融-国学对应:Tested in Practice, and Marked When Not(验过的和听来的,要在同一张纸上分开写)↔ 验之行事
段一·现身
十一月二十七日,早上七点。
马可把那个纸箱拆开了。
里面两个塑料袋,用绳子扎着。
一袋杏干,一袋葡萄干。
都是两斤。
他把每一袋分成两半。
一半装进小袋,扎紧,做记号。
这一半不开,是对照。
另一半摊在窗台上,敞着口。
窗台朝东,早上有太阳,下午没有。
他从抽屉里找出一个厨房秤。
是去年买的,称面粉用的,精到一克。
他把四个袋子都称了一遍。
记在本子上。
第一天:
杏干开口 五百零三克。
杏干封口 五百零一克。
葡萄干开口 四百九十八克。
葡萄干封口 五百克。
他又找了一个温湿度计,是网上买的,三十七块,第二天到。
在它到之前,他用手机上的天气看湿度。
十一月二十七日,成都,湿度百分之七十四。
七点二十,他吃了第一个杏干。
颜色是深褐的,几乎发黑。
表面是干的,捏上去有一点硬。
放到嘴里,先是酸,然后回甜。
很酸,比他想的酸。
咬到最后有一点点渣。
他在本子上写:
「第一天。硬。酸。回甜。不粘手。」
写完他看了一眼这一行。
十二个字。
他想了一下,把「不粘手」三个字画了一个圈。
因为这三个字是这次要看的那三个字。
段二·古文镜像
“作乾蒲萄法:极熟者一一零叠取,刀子切去蔕,勿令汁出。蜜两分,脂一分,和内蒲萄中,煮四五沸,漉出,阴乾便成矣。非直滋味倍胜,又得夏暑不败坏也。”
“采捃经传,爰及歌谣,询之老成,验之行事,直取会意,不尚浮辞。”
English Mirror: Method for making dried grapes: take only the fully ripe, one by one; cut away the stem with a knife and do not let the juice run… strain them out and dry them in shade, and they are made. Not only is the flavour twice as good, they also keep through the summer heat. — Gathered from the classics and commentaries, and from the songs of the people; asked of those with experience; and tested in practice. Take the sense only, and do not prize fine words.
那件事,是马可在第三天晚上查到的。
他当时是在找一个别的东西。
他想知道葡萄干到底能放多久,想找一个权威一点的说法。
结果他搜到了《齐民要术》。
他知道这本书,可是没有读过。
他找到了那一段,是卷四里的。
讲怎么做干葡萄。
他一句一句对着白话看。
看到最后四个字的时候,他坐直了。
「阴乾便成矣。」
阴干。
一千五百年前这本书上写的是阴干。
马可把手机放下。
常玉山那句「不见太阳」,写在这儿了。
写在公元六世纪。
他往下又看了一句。
「非直滋味倍胜,又得夏暑不败坏也。」
不光味道好,还能过夏天不坏。
马可在本子上把这一句抄下来。
抄完他在旁边写了一行:
「一千五百年前,人家关心的也是能放多久。」
第三天晚上,马可把《齐民要术》的序看了一遍。
他本来只想看那一段做法。
序他是顺手点开的。
里面有一句,他看了三遍。
「采捃经传,爰及歌谣,询之老成,验之行事。」
前面三句他懂:从书上找,从民间的歌谣里找,问有经验的老人。
第四句他停住了。
验之行事。
试过。
马可在本子上写了一行:
「他找了三种来源,最后一条是自己试。」
他停了一下,写了下一行:
「我这十七天,前三种都做了。」
又写了一行:
「第四种是从今天早上七点开始的。」
写完这三行,他又坐了一会儿。
他想的是一件跟葡萄干没有关系的事。
他在纽约做了六年。
那六年他做的是第一条:采捃经传。
看年报,看招股书,看行业研报。
偶尔做第二条:听市场上的传闻。
第三条他做得很少。
打电话给行业里的人,一年也就十几回。
第四条,一次也没有。
他从来没有去过一家他分析过的公司的车间。
马可在本子上写了一行:
「六年,我做的是四分之一。」
他停了一下,把这句改准了:
「六年,我做的是那四条里最便宜的一条。」
他又想了想,在下面加了一行:
「最便宜的那一条,也是最多人做的那一条。」
再加一行:
「所以那一条上,没有便宜可占。」
他又在下面写了一行,这一行是问自己的:
「那第四条为什么没人做?」
他想了几分钟。
答案很简单,简单到有点让人不高兴。
因为第四条慢。
看一份年报,两个小时。
打十个电话,一天。
称二十天,二十天。
而且二十天以后,得到的可能是「没变化」。
他在本子上写:
「第四条最贵的不是钱,是它可能什么也得不到。」
写完他看了一眼窗台上那两袋。
今天是第三天。
什么也没有发生。
他忽然想到,要是二十天什么也不发生,他今天写的这些就全白写了。
他又想了一下,觉得不对。
要是二十天什么也不发生,那也是一条。
那一条是:成都敞着放二十天没事。
这一条要是真的,它值多少?
他算了一下。
它值的是三月那两百袋可以敞着发。
省下每一袋五毛五。
两百袋一百一十块。
一百一十块。
马可看着这个数,笑了一下。
二十天,一百一十块。
他在本子上写了一行:
「一百一十块,二十天。」
他停了一下,加了一句:
「可是这一条今年值一百一,明年值一千一,后年值一万一。」
又加了一句:
「因为它是一条,不是一次。」
这一行写完他把笔停住了。
因为这句话解释了他这两年一直没想通的一件事。
他在纽约的时候,看的报表跟别人看的是同一份。
同一份报表,一万个人在看。
他能比别人多看出什么?
答案是几乎没有。
十七天前他在库尔勒问的第一个人,是蒋姐。
那张表上二十一个群,三月到六月是空的。
那张表,全世界只有一个人给过他。
段三·事件主体
第九天,十二月五日。
这九天什么也没有发生。
杏干还是硬的。
葡萄干还是那个颜色。
马可每天早上称一次,吃一个,写四五个字。
他有两天差点忘了。
第九天早上,秤上的数变了。
杏干开口那一袋:五百零六克。
第一天是五百零三。
多了三克。
三克,百分之零点六。
马可站在秤前面,看了一会儿。
三克这个数太小了。
小到可能是秤不准。
他把那袋拿下来,又放上去。
还是五百零六。
他把封口那一袋也称了。
五百零一。
跟第一天一样。
马可在本子上写了一行,写完在旁边画了一个三角:
「第九天。开口+3克。封口不变。」
他把常玉山那句想了一遍。
返潮先是长分量。
等你尝出来,已经晚十天了。
他当时觉得这句是经验之谈。
今天早上他知道这句是仪器。
第十二天,十二月八日。
杏干开口:五百一十克。
七克。
百分之一点四。
这天早上他吃那一个的时候,觉得不一样了。
不是味道。
是手感。
从袋子里拿出来的时候,有两个是连在一起的。
不是粘死,是轻轻一碰就分开。
可是第一天不是这样。
第一天倒出来是散的。
马可在本子上写:
「第十二天。+7克。两个粘在一起,一碰就开。」
他把这一条拍了照,发给常玉山。
常玉山回得很快:
「第十二天,早了。」
下面一条:
「我这儿要到第七十天。」
马可看着这两条,在屋里坐了很久。
七十天对十二天。
差六倍。
那天晚上,马可做了一件事。
他给郭老师打了个电话。
他说明天想再去一趟,带点东西。
第二天他去了,带了那半袋封着的葡萄干。
他跟郭老师说了这件事。
说完他问:
“郭老师,您能不能帮我放一份在车库里。”
郭老师说:
“放着就行了?”
马可说:
“还要每天称一次。”
他把秤的事说了。
郭老师听完,问了一句:
“要不要写下来?”
马可说:
“要。”
郭老师说:
“那我记在本子上。”
她把那个红皮本拿出来,翻到后面空白的一页。
她在最上面写了一行:葡萄干。
下面写了日期。
马可站在那个车库里看着她写。
他忽然觉得这件事有点奇怪。
一个成都的退休老师,在自己家车库里,给一个敦煌农民的货称重量。
记在一个跟敦煌那个村里一模一样的本子上。
中间三千公里。
两个人不认识。
那半个月,两个地方的数不一样。
马可的窗台:第十二天七克,第十七天十四克。
郭老师的车库:第十二天四克,第十七天九克。
同一个城市,同一批货,差得很明显。
马可去看了两次,才找出来为什么。
第一,郭老师那个车库温度低。
他的窗台白天二十二三度,车库十八度。
第二,也是更要紧的一条。
郭老师的袋子放在货架上,离地七十公分。
下面还垫了一块防潮垫。
马可的袋子放在窗台上。
窗台底下是暖气片。
他每天早上开窗通风,外面正在下雨。
马可在本子上写了一行:
「同一个城市,两间屋子,差五天。」
他停了一下,加了一句:
「我要是只放一份,我今天算出来的数是错的。」
他又加了一句:
「而且我不会知道它错。」
他把这三行写完,又想了一件事。
要是他有三个地方呢?
他掰着手指算了一下能找谁。
陈姐。她那儿是一个仓,租的,在三环外。
小雷。他没有点。
蒋姐。她的货是当天到当天分,不存。
算下来,成都这么大,他能放东西的地方只有两个。
一个是他自己的窗台。
一个是郭老师的车库。
马可在本子上写了一行:
「我在成都两年,能放一袋东西的地方,只有两个。」
他盯着这一行看了很久。
因为这一行说的不是储存。
这一行说的是,他在这个城市里,能麻烦的人只有一个。
他把这句写下来了:
「郭老师是我在成都唯一能说『您帮我称一下』的人。」
写完他又加了一句:
「而我认识她第九天。」
贾思勰是第十七天晚上出现的。
那天是十二月十三日。
马可刚称完,写完,坐在桌前。
杏干开口那一袋:五百一十七克。
十四克。
百分之二点八。
他把袋子倒出来一点在手心里。
这一次不是两个连在一起。
是一整片。
倒不出来,得抠。
他捏了一下,两个杏干粘在一起,没有分开。
他试了三次,最后是用手掰开的。
掰开的地方拉出丝来。
就在这个时候,桌子对面坐了一个人。
穿的是很旧的官服,颜色说不清。
五十来岁,手很粗,不像坐办公室的。
指甲缝里有黑的。
他说的第一句是:
“你记了十七天。”
马可说:
“是。”
那个人说:
“贾思勰。”
那天晚上,贾思勰在那张桌子前面,翻了马可的本子。
他翻得很快。
翻到第九天那一页,他停了一下。
他指着那个三角,问:
“这个是什么?”
马可说:
“这一天变了。”
贾思勰说:
“变了三克。”
马可说:
“是。”
贾思勰把本子推回来。
他说了一句,马可当天晚上抄了三遍:
“三克这一笔,比后头那十四克值钱。”
马可问:
“为什么?”
贾思勰说:
“十四克,谁都看得出来。”
他把手放在桌上。
“三克,只有称过的人知道。”
那天晚上,马可问了那个他这两天一直在想的问题。
“先生,您那本书里,写了多少是您自己试过的?”
贾思勰说:
“不多。”
这个回答马可没想到。
他问:
“那剩下的呢?”
贾思勰说:
“书上抄的,老农说的,歌谣里唱的。”
马可问:
“那怎么分?”
贾思勰说:
“我注了。”
他说得很平常。
他说:
“从书上抄的,我写出处。”
“老农说的,我写谁说的。”
“我自己试过的,我写试的时候是什么天。”
马可问:
“没试过又不知道从哪儿来的呢?”
贾思勰说:
“那我写未试。”
马可在那张桌子前面坐着,把这句在心里过了两遍。
他问:
“先生,写未试,不是自己打自己的脸吗?”
贾思勰说:
“打脸。”
他承认得很干脆。
然后他说了一句,这一句是这一回的骨:
“可是不写未试,后头的人试出来是错的,整本书都不能信了。”
他把手抬了一下。
“写了未试,错的只是那一条。”
马可拿起笔,把这两句写下来。
写完他抬起头。
他问:
“先生,那我三月那七行字,怎么办?”
贾思勰问:
“哪七行?”
马可念了一遍。
念完贾思勰问了一句:
“这七行,哪几行你试过?”
马可想了一下。
他说:
“四斤半出一斤,我没试过。”
他接着说:
“四十五天,我没试过。”
“不见太阳,我进过那个房子,可是我没有见过挂的时候。”
他停住了。
七行里,他自己验过的,一行也没有。
全是听来的。
贾思勰说:
“那你写听谁说的。”
马可问:
“只写这个?”
贾思勰说:
“再加一行你自己试过的。”
他站起来。
他说了走之前那一句:
“你今天有一条了。”
马可站起来,追问了一句:
“先生,一条够吗?”
贾思勰已经走到门口了。
他回过头。
他说:
“一条不够。”
马可问:
“那要几条?”
贾思勰说:
“我这一辈子,攒了不到二百条。”
马可算了一下。
《齐民要术》十卷九十二篇。
里头几千条法子。
自己试过的不到二百条。
他问:
“先生,那您写那本书,图什么?”
贾思勰说:
“我图的是不散。”
马可没听懂。
贾思勰说:
“这些法子在老农嘴里。”
他抬起手。
“一个老农死了,他那几条就没了。”
他说:
“我把它写下来,它就不跟着人死。”
马可站在那儿。
他问:
“可是没试过的写下来,万一是错的呢?”
贾思勰说了一句,说完就不在了:
“错的留在纸上,将来有人能改。”
他的声音已经很远:
“没写下来的,将来没有人知道有过。”
段四·钩子结尾
第二十天,十二月十六日。
那天早上马可把两袋都倒出来了。
杏干开口那一袋,已经不能叫一袋杏干。
是一块。
五百二十一克,比第一天多十八克。
掰开还能吃,可是没有人会想吃。
葡萄干开口那一袋:五百零九克,多十一克。
倒出来还是散的,可是捏在手里有一点黏。
颜色比第一天暗,能看出来。
两袋封口的:五百零一,五百。
跟第一天一模一样。
郭老师那边,杏干多十二克,葡萄干多七克。
马可把这四组数,抄在一张干净的纸上。
抄完他在下面写了三行:
一,敞开放,成都,十七天开始粘。
二,封着放,成都,二十天没有变化。
三,问题不在成都,在开口。
他把这三行看了很久。
因为这三行的意思是,这件事不用冷库。
只要袋子好。
那天上午,马可算了一笔账。
原来的袋子是普通透明袋,八毛。
他查了三种。
第一种,真空。四毛加钱,可是抽完真空葡萄干压扁了,难看。
第二种,加干燥剂。一毛五,可是袋子里多一小包东西,有人会问。
第三种,铝箔袋。两毛五,可是看不见里头的绿。
绿是这个东西唯一好看的地方。
马可在纸上把这三种写完,坐了半天。
然后他想了第四种。
外面一个铝箔袋。
里面一个透明小袋。
拆开外面那层,才看得见绿。
他算了一下:两个袋子加起来,一块三五。
比原来多五毛五。
一袋毛利从十一块三,降到十块七五。
两百袋,少一百一。
他把这个数写下来,没有画圈。
因为这个不用犹豫。
他在旁边写了一行:
「五毛五,买的是第十七天。」
那天下午,他给常玉山发了一长条。
他把二十天的数全发过去了。
两个地方,四组,每天一个数。
发完他加了一句:
「常师傅,您那个六个月是对的。在敦煌是对的。」
下面一句:
「在成都,敞着放是十七天。」
这两条发出去,那边很久没有回。
过了大概四十分钟,常玉山回了。
回的不是字,是一张照片。
照片拍的是一个本子。
很旧的本子,纸都黄了。
上面是手写的,一行一行,全是数。
马可把照片放大。
他看清了最上面那一行的日期。
二〇〇三年八月。
下面一行一行,每一行两个数。
一天两次。
从八月记到十二月。
马可数了一下,一百多天。
照片下面,常玉山打了一句:
「这是零三年那次。」
过了一会儿,又一条:
「我记了一百二十六天,最后没弄明白。」
再一条:
「你二十天弄明白了。」
马可看着这三条,在屋里站着。
他打了一句,打完停了很久才发:
「常师傅,我没有弄明白。」
下面一条:
「我只是有两个地方,您只有一个。」
这一条发出去,那边过了十分钟。
回来的只有六个字:
「明年我放两个。」
马可看着这六个字,忽然明白了一件事。
他这两天算的所有账,都是袋子和毛利。
可是刚才那一条过去,改的不是袋子。
是常玉山明年怎么做事。
他在本子上写了这一天最后一行:
「我给他的不是订单。」
「是一个他自己那儿没有的第二个地方。」
写完这两行,他把本子往前翻。
翻到十一月十三日,酒泉那一页。
那一页上写着霍去病那句:一坛酒倒进一口泉,什么味也没有。
旁边是他自己写的:我不是要他们喝到,我是要他们知道有。
马可在这一页下面,补了一行今天的字:
「一个第二地点,也是这样。」
他停了一下,写完:
「常师傅不需要我给他钱。」
「他需要知道,有一个地方,跟他这儿不一样。」
他把笔放下,忽然想到一个很实际的问题。
明年他要放两个地方。
放在哪儿?
敦煌那六座房,是同一块地上的六座。
放在第一座和第六座,天气是一样的。
他打了一句过去:
「常师傅,您明年那两个放哪儿?」
这一条发出去,那边回得很快:
「一个我这儿,一个吐鲁番。」
马可看着这一行,坐直了。
因为这意思是,常玉山已经跟那边通过气了。
他打了一句:
「您跟他说了?」
常玉山回:
「说了。」
下面一条:
「他说他那边一百多座房,从来没有人往外称过。」
再一条:
「他说他愿意。」
马可把这三条看了两遍。
吐鲁番那个人,一百多座房。
在这条线上,他是最大的一个。
这个人愿意做的事,是每天称一次一袋葡萄干,然后把数发给一个成都的年轻人。
马可在本子上写了一行:
「最大的那一个,愿意做最小的那一件。」
他停了一下,写完:
「因为这件事,他那儿也没有人做过。」
欲知后事如何,请看下回,《二月二十号那三十袋》
附录一·Dongpo's Note
27 November, seven a.m. Opened the carton. Two bags tied with string, two jin each, apricot and raisin. Split each in half: one half into a small bag, sealed and marked, the control; the other half open on the east windowsill. Kitchen scale from last year, accurate to the gram.
Day one: apricot open 503 g, sealed 501; raisin open 498, sealed 500. Chengdu humidity that day, 74%.
Seven twenty, the first apricot. Dark brown, nearly black; dry to the touch, slightly hard; sour first, then sweet coming back, and sourer than I expected; a little grit at the end. Day one. Hard. Sour. Sweet after. Not tacky. Twelve characters, and I drew a circle round the last three, because those are the three the whole thing is about.
On the third night, looking for something else — how long raisins actually keep, from some authority — I found the Qimin Yaoshu, a book I knew of and had never read. Book four, the method for dried grapes. Reading it against the modern gloss, I sat up at the last four characters. Dry them in shade, and they are made.
Chang's no sun gets in, written down in the sixth century. And the line after it: not only is the flavour twice as good, they also keep through the summer heat. Fifteen hundred years ago the question was already how long it lasts.
Then the preface, opened by accident, one sentence read three times. Gathered from the classics and commentaries, and from the songs of the people; asked of those with experience; and tested in practice. The first three I understood. The fourth stopped me. Tested in practice. In seventeen days out west I had done the first three. The fourth began at seven that morning.
Day nine, 5 December. Nothing at all for nine days, and I nearly forgot twice. Then the scale: apricot open 506, three grams up. Three grams is small enough to be the instrument, so I took it off and put it back. Still 506. Sealed bag: 501, unchanged.
Which is Chang's damp shows up as weight first; by the time you can taste it you're ten days late. I had filed that as folk wisdom. It is an instrument.
Day twelve: 510, up seven, 1.4%. And the first change you could feel rather than weigh — two of them came out of the bag joined, parting at a touch, where on day one they poured out loose. Sent it to Chang. Day twelve is early. And: here it takes seventy days.
Seventy against twelve. Six times.
So I took the sealed half to Guo and asked her to run one in her garage, weighed daily. Just leave it there? No — weighed every day. She said: should I write it down? Yes. Then I'll put it in the book — the red one, at a blank page near the back, "raisins" at the top and the date under it. A retired teacher in Chengdu weighing a Dunhuang farmer's fruit into a notebook identical to the one in that village, three thousand kilometres apart, the two of them strangers.
The two sites diverged. My sill: +7 at day twelve, +14 at seventeen. Her garage: +4 and +9. Same city, same batch. It took two visits to find why: her garage runs 18° against my 22–23, and her bags sit on a rack seventy centimetres off the floor on a damp mat, while mine sat on a windowsill above a radiator, next to a window I open every morning onto rain.
Same city, two rooms, five days apart. If I had run one sample, today's number would be wrong — and I would not know it was wrong.
Day seventeen, 13 December. Apricot 517, +14, 2.8%. Not two stuck together now; the whole thing comes out as a slab, and has to be prised. I pulled two apart and they drew threads.
And a man was sitting on the other side of the table in a very old official's robe of no determinable colour, fifty-odd, with coarse hands and black under the nails, nothing like someone who sits at a desk. You have kept this seventeen days.
He went through the notebook fast, stopped at day nine, pointed at the triangle. What's this? That's the day it changed. It changed by three grams. Yes. He pushed the book back. That three-gram entry is worth more than the fourteen grams later. Why? Anybody can see fourteen grams. And, hand flat on the table: only a man who weighed it knows about three.
Then the question I had been carrying. How much of your book did you test yourself? Not much. I had not expected that. And the rest? Copied from books, told to me by old farmers, sung in ballads. How does a reader tell? I annotated it. Copied: he gives the source. Told: he gives the man. Tested: he records what the weather was when he tested it. And what he neither tested nor could source? I write: not tried.
Isn't writing "not tried" a slap in your own face? It is. No hedging at all. And then: but if I don't write it, and someone later finds that one wrong, the whole book becomes unusable. If I write it, only that line is wrong.
So what about my seven lines for March? Which of them have you tested? Four and a half to one — not tested. Forty-five days — not tested. No sun — I have stood in the building but never seen it hung. Of seven lines I had verified none. All of it heard.
Then write who told you. And? Then add one line you tested yourself. He stood up. You have one now. — M.
附录二·Marco's Memo
Memo to file. Chengdu, 16 December 2027. Twenty-day storage trial, closed.
1. Design. Two products (apricot, raisin), two treatments (open / sealed control), two sites (my windowsill, 22–23°C; Guo's garage, 18°C, racked 70 cm off the floor on a damp mat). Daily gravimetric measurement to 1 g plus a sensory note. Chengdu RH 74–78% throughout.
2. Results (open bags, gain from day 1). — Windowsill: d9 +3 g (0.6%) / d12 +7 (1.4%) / d17 +14 (2.8%) / d20 apricot +18, raisin +11. — Garage: d12 +4 / d17 +9 / d20 apricot +12, raisin +7. — Sealed controls at day 20: 501 and 500 g against 501 and 500 on day one. No measurable change. — Sensory: first adhesion at d12 (two pieces joined, parting at a touch); d17 apricot consolidated into a slab, drawing threads on separation; d20 raisin tacky and visibly darker but still loose.
3. Conclusions. (a) Chang's six months is correct — in Dunhuang. At 30% RH it is presumably conservative. At 74–78% RH, open, the useful life is seventeen days. A factor of about six against his stated seventy. (b) The failure is not the city. It is the open bag. Sealed product was unchanged at day twenty in the same room. Therefore no cold store, no cold chain, and no change to the freight model. The fix is packaging. (c) Two sites in one city differed by five days, driven by temperature (18 vs 22°C) and by floor proximity. A single-site trial would have produced a confidently wrong number with no indication that it was wrong. File next to ch.132: replication without decorrelation is not replication — and its converse, which I now have from the other side: two samples in genuinely different conditions bought me the error bar, which one sample cannot do at any sample size.
4. The leading indicator, and the reason to keep weighing. Chang: damp shows as weight first; by the time you taste it you're ten days late. Confirmed exactly: first weight signal d9, first tactile signal d12, first signal a customer would notice d17. Eight days of warning, available only to someone holding a ¥40 kitchen scale. Jia Sixie's version: anybody can see fourteen grams; only a man who weighed it knows about three — the three-gram entry is worth more than the fourteen.
5. Packaging decision. — Current: clear PE, 0.80. — Vacuum: +0.40, but compresses the fruit and looks bad. Rejected. — Desiccant sachet: +0.15, but an unexplained sachet inside food invites questions. Rejected. — Foil: +0.25, but kills the visible green, which is the only visually attractive property the product has and the thing that distinguishes it from brown (ch.129: green is worth double). — Adopted: outer foil, inner clear. 1.35 total, +0.55. Gross per bag 11.30 → 10.75; on 200 bags, −¥110. Not a marginal call. Fifty-five fen buys day seventeen. And the outer bag carries the seven lines, so the green is not seen until the buyer opens it — which is ch.134's third rule made physical: an opaque surface that has to be opened is asked about; a transparent one is glanced at. — Additional, and it belongs on the bag rather than in my head: "eat within ten days of opening." Stating the limitation is the Dingcheng brick again (ch.127): a buffer in plain sight is the calculation; a buffer hidden is an error.
6. The provenance rule for the March copy, adopted from Jia Sixie and this is the most portable thing in this entry. His book distinguishes four kinds of claim and marks each: copied (source given) / told (informant named) / tested (conditions recorded) / not tried (said so). Audit of my seven lines: 4.5:1 — heard. Forty-five days — heard. No sun — heard, building seen, hanging never seen. Six houses — seen. Age sixty-two — asked. Verified by me: none. Rewriting accordingly: attribute what is heard ("Chang Yushan says"), and add one line of my own that I have tested. That line now exists: "Sealed, twenty days in Chengdu, no change. Open, it starts to stick around day seventeen. I ran it in two places from 27 November. Eat within ten days of opening." Why write "not tried" against my own claims, when it is self-inflicted damage: if I don't mark it and one line proves wrong, nothing else I have written can be trusted. If I do, only that line is wrong.
7. The exchange that mattered more than the trial. Sent Chang all four series. He replied with a photograph of a yellowed notebook: August 2003, two readings a day, a hundred and twenty-six days. The year one whole house went sticky on him. I kept it a hundred and twenty-six days and never worked it out. You worked it out in twenty. Which is not true and I said so: I didn't work it out. I just had two places and you had one. Ten minutes, then six characters: next year I'll run two. I have been costing bags and margins all week. That message changed how a man will work next season. The order is 200 bags and ¥5,200. What I actually supplied was a second site, which he could not obtain from where he stands.
8. Open. Confirm the foil/clear laminate cost at 200-unit quantity (quoted, not sourced). Guo's garage becomes the standing Chengdu test site — ask her terms; she should not do this for nothing (ch.131: the ledger-keeper is unpaid but gets the best table, and I do not have a table to give). And run the same trial on the raisin in March conditions, not December: RH is comparable but temperature will be higher.
附录三·English Highlight
"Strain them out and dry them in shade, and they are made. Not only is the flavour twice as good, they also keep through the summer heat."
This is the method for dried grapes as set down in the sixth century, and the operative words are the last four: dry them in shade. A farmer outside Dunhuang, asked why a wall carries four hundred slanted holes, gives the same answer in four characters: no sun gets in. Fifteen hundred years, and the instruction has not changed, nor has the reason for it, nor has the question the passage ends on — how long will it keep.
The preface to the same book contains a sentence worth more than the recipes. Gathered from the classics and commentaries, and from the songs of the people; asked of those with experience; and tested in practice.
Three of those are collection. The fourth is different in kind, and it was the fourth that produced the only useful fact of a twenty-day experiment involving two jin of dried fruit, a kitchen scale and a windowsill.
Nothing happened for nine days. On the ninth the scale read three grams heavier. Three grams is small enough to be the instrument, so it was taken off and put back. On the twelfth, seven grams, and the first two pieces came out of the bag joined. On the seventeenth, fourteen, and the whole thing had to be prised apart in a slab. The sealed control, sitting in the same room, was unchanged on day twenty.
The producer had said the product held six months. It does — where he lives, in air with a third as much water in it. Where it is eaten, open, it holds seventeen days.
And when the notebook was read by someone qualified to read it, the entry he stopped at was not the fourteen grams. That three-gram entry is worth more than the fourteen grams later. Anybody can see fourteen grams. Only a man who weighed it knows about three.
Then the harder question, put to a man who spent his life compiling agricultural methods: how much of your book did you test yourself? Not much. And the rest? Copied, or told, or sung. How is a reader to know which is which? I annotated it — the source for what was copied, the informant for what was told, the weather for what was tested, and for the remainder, three characters: not tried.
Isn't that a slap in your own face? It is. And then, with no hedging at all: but if I don't write it, and someone later finds that one line wrong, the whole book becomes unusable. If I do write it, only that line is wrong.
附录四·Dongpo's Cheat Sheet
| 二十天 | 出处 | 现代对应 | 一句话 |
|---|---|---|---|
| 一分为二 | 一半封口做对照 | 一半敞在窗台 | 厨房秤,精到一克 |
| 第一天十二个字 | 硬·酸·回甜·不粘手 | 「不粘手」画了圈 | 这三个字是这次要看的 |
| 阴乾便成矣 | 《齐民要术》卷四 | 常玉山「不见太阳」 | 写在公元六世纪 |
| 采捃经传,爰及歌谣,询之老成 | 前三条都是收集 | 验之行事 | 第四条从今天早上七点开始 |
| 第九天+3克 | 秤上第一个信号 | 拿下来又放上去,还是506 | 封口那袋不变 |
| 返潮先是长分量 | 常玉山那句 | 原以为是经验之谈 | 其实是仪器 |
| 第十二天+7克 | 两个粘在一起,一碰就开 | 常玉山:第十二天,早了 | 我这儿要到第七十天 |
| 两个地方 | 窗台22-23度/车库18度 | 车库离地七十公分加防潮垫 | 窗台底下是暖气片 |
| 同一个城市,两间屋子,差五天 | 只放一份,算出来的数是错的 | 而且不会知道它错 | |
| 第十七天 | +14克,一整片,掰开拉丝 | 封口那两袋二十天没变 | 问题不在成都,在开口 |
| 不用冷库 | 只要袋子好 | 运费模型不动 | |
| 三克那一笔比十四克值钱 | 贾思勰 | 十四克谁都看得出 | 三克只有称过的人知道 |
| 您书里多少是自己试过的 | 不多 | 抄的写出处,听的写谁说的 | 试过的写试的时候是什么天 |
| 没试过的写「未试」 | 这不是自己打脸吗 | 打脸 | 不写,整本书都不能信;写了,错的只是那一条 |
| 七行里我验过几行 | 一行也没有 | 全是听来的 | 那你写听谁说的,再加一行你自己试过的 |
| 外铝箔内透明 | 一块三五,比原来多五毛五 | 一袋毛利11.30→10.75 | 五毛五,买的是第十七天 |
| 开封后十天内吃完 | 印在袋上 | 与定城砖同 | |
| 零三年那本 | 一天两次,一百二十六天 | 我记了一百二十六天没弄明白 | 你二十天弄明白了 |
| 我没有弄明白 | 我只是有两个地方 | 您只有一个 | 明年我放两个 |
| 我给他的不是订单 | 是一个他那儿没有的第二个地方 |