从希腊罗马化外探索叙述看西方古代纪实文类

一 · 接收矩阵：宿主如何处理无从核实的报告1 · Reception Matrix: how hosts handle reports they cannot verify

行＝有真实宣称的宿主作者，列＝探险家（其文本多已佚，仅存于宿主体内）。格色＝姿态。点击任一格读原文。Rows = host authors who make a truth-claim; columns = explorers (whose texts mostly survive only inside their hosts). Cell colour = stance. Click any cell to read the source.

二 · 反讽之链2 · The Irony Chain

一条自我吞噬的链：怀疑论内在于纪实文类的规范意识本身。A self-devouring chain: skepticism is internal to the genre’s own norm of truthfulness.

三 · 可信度的悖论：亲见 ≠ 真实3 · The Credibility Paradox: autopsy ≠ truth

横轴：亲见 → 传闻 → 奇物；纵轴：采信 → 驳斥/切割。悖论（◎）：Patrocles 亲测里海却错，Herodotus 扶手椅反而对。Horizontal: autopsy → report → marvel; vertical: relay → refute/sever. The paradox (◎): Patrocles measured the Caspian first-hand and was wrong; Herodotus, from his armchair, was right.

四 · 冒险的产出：真值谱系4 · The Payoff of Risk: a Truth-Valency Spectrum

被琉善一概斥为谎言的奇观，未必尽属编造。以真值衡量，其中若干藏着可核的内核：古代记述在左，近世指认在右。古代 locus 已逐字亲核；现代指认属科学史／接收史判断，一律以“近世或以为”出之，非古人自指。The marvels Lucian would dismiss wholesale are not all fabrication. Measured for truth-valency, several conceal a verifiable kernel: the ancient report on the left, the modern identification on the right. The ancient loci are verbatim-checked; the modern identifications belong to the history of science and of reception — given throughout as “modern scholarship has argued”, never as an ancient self-identification.

五 · 语境5 · Context

上方任一格 / 链接点开后，逐字原文在此。Open any cell or chain-link above; the verbatim source appears here.

点击上方矩阵格或反讽链，阅读逐字希腊/拉丁/古汉语原文。Click a matrix cell or a chain-link above to read the verbatim Greek/Latin/Chinese source.

六 · 行程路线复原图6 · Route reconstructions

全部条行程的路线复原。沿岸段为 10 角分地形栅格上的近岸最小成本路径（示意，非实测）；外海段为最小成本水路；河流段沿 Natural Earth 河道中线；细点段为转述或推断、未经亲证。选择一条行程以突出显示；编号圆点标出该行程的既定途经点（重复途经点如实重复，例如埃利乌斯·加卢斯经奈季兰南下马里卜、复经奈季兰折返）。Route reconstructions for all — expeditions. Coasting legs are least-cost offshore-band paths on a 10-arc-minute relief grid (indicative, not surveyed); open-sea legs are least-cost water paths; river legs follow Natural Earth centerlines; faint dotted legs are reported/inferred, not attested. Choose an expedition to highlight it; numbered pins mark its placed waypoints (repeated waypoints repeat honestly — e.g. Aelius Gallus passes Najran outbound to Marib and again on the retreat).

━ 沿岸coast · ╴╴ 外海open sea · ···· 陆路overland · ╴· 河流river

七 · 比较项：甘英，公元九七年7 · A comparator: Gan Ying, AD 97

本录诸条记录大半取自希腊罗马文献一系；其中三条出《史记》，为已标明的越界扩充，即张骞使西域、徐福入海求仙、汉通西南夷。永元九年班超遣甘英使大秦，抵条支，临大海将渡，闻安息西界船人言海道之险而止。以本录他处所求者衡之，有名有姓的使者，有年可系的差遣，有站可循的行程，有据可查的折返，四者具备；公元三百年以前，欧亚两端记载之完备，无过于此一次探路者。其所以不列为本录之一条，正因为这一比较的力量来自它站在框外。The records in this reader are drawn largely from the Greco-Roman documentary tradition; three of them rest on the Shiji and enter as flagged extensions of it — Zhang Qian, Xu Fu’s search at sea, and the Han probes toward the southwest. In AD 97 the protector-general Ban Chao sent Gan Ying toward Da Qin; he reached Tiaozhi at a great sea and turned back on hearing the Parthian frontier sailors describe the crossing. Measured by what the reader elsewhere requires — a named agent, a dated commission, a station itinerary, a documented turn-back — this is the best-attested reconnaissance of the pre-300 CE corpus at either end of Eurasia. It is held here as a comparator rather than as a record of its own, because the comparison does its work precisely by standing outside the frame.

选择上方地图中的“◇ 甘英使大秦”一项，可见其形：可靠落点唯有出发地一处；入安息一段作转述／推断的细点线；条支、阿蛮、斯宾、于罗四站一律未定位，不予落点；海亦未予指认，故不绘。斯宾之为泰西封、于罗之为希拉，出张星烺按语，属编者推断，非《后汉书》原文。Choose “◇ 甘英使大秦” in the map above and the shape is plain: one securely pinnable point, the departure. The traverse into Parthia is drawn as a reported/inferred leg. Tiaozhi, Aman, Sibin and Yuluo are all unlocated and left unpinned; the sea is unidentified, so no sea is drawn. The equations of Sibin with Ctesiphon and Yuluo with Hira come from Zhang Xinglang's editorial notes, not from the Hou Han shu.

两端之不对称，所系者在文献与今人考订，而不在古人足迹之远近。希腊罗马一系地名之所以密可落点，赖有长期的校勘传统与今日的地名权威库；《后汉书》西行诸站则仅存于后世汇编的转录，其指认至今互竞。故本录呈中海之形，所记者乃“何种记载得以传世并得以定位”，而甘英一线之几乎不可落点，本身即为此点之证。The asymmetry between the two ends belongs to the documentation and to modern scholarship rather than to how far ancient people travelled. The Greco-Roman toponyms are densely pinnable because of a long editorial tradition and today's gazetteer authorities; the western stations of the Hou Han shu survive through a later compilation's relay, and their identifications remain in competition. The Mediterranean shape of this reader therefore records which accounts survived and became locatable — and the near-total unpinnability of Gan Ying's line is itself the evidence for that.

证据：《后汉书》卷八八，据张星烺《中西交通史料汇编》第一册（2003年版）页一六一至一六二转录，标为“经汇编转引”，俟核于校勘本；西行地理之框架承自《汉书》卷九六上（中华书局点校本，页三八八九“南道极矣，转北而东得安息”“西与条支接”），故两史地名相同处不得计为各自独立之证。Evidence: Hou Han shu 88, relayed by Zhang Xinglang, Zhongxi jiaotong shiliao huibian I (2003 ed.), 161–162, flagged as later-relay-via-compilation pending verification against a critical edition. The western geography is inherited from Han shu 96A (Zhonghua ed., 3889: the south road ends, one turns north and then east to reach Parthia; Tiaozhi lies to its west) — so toponyms shared by the two histories do not count as two independent attestations.