数据错误 / Data error: 陳文龍 (15213) coded as grandson (孫) of 陳俊卿 (10884); sources say 之後／五世從孫 (collateral)

[reggiechan74]

Thanks for maintaining CBDB and the SQLite distribution. This is a kinship-coding error in the bulk SQLite DB; I've traced it back to the cited source and reconstructed the lineage so it should be quick to verify.

中文 TL;DR. CBDB 现将 陳文龍 (15213) 记为 陳俊卿 (10884) 之孫（KIN_DATA kincode 243 SS），反向记为「祖父」（kincode 62 FF，来源＝「未知」0）。然而所引来源《宋人傳記資料索引》(source 7596) 所本之 《宋史·卷451·陳文龍傳》 原文作「丞相俊卿之後也」（後 = 后裔，非孫）；學界與玉湖陳氏族譜均作「五世從孫」（旁支），其父為 陳粢。文龍與陳俊卿分屬玉湖陳氏自二世祖貴以下的兩支：俊卿出詵房（四世），文龍出德戴房（八世）。建議將此祖孫關係改為旁系/遠裔或刪除待核，並補錄其父陳粢；文龍與陳瓚 (15201) 之叔姪關係有《宋史》為據，應保留。

English summary. CBDB encodes a lineal grandfather→grandson link between Junqing (10884) and Wenlong (15213). The primary source the citation rests on says only descendant (之後), not grandson (孫); scholarship and the 玉湖陳氏 genealogy read 五世從孫 — a collateral descendant. The two men sit on two different branches of the Yuhu Chen that split at the 2nd-generation ancestor 貴.

1. The records (bulk SQLite KIN_DATA)

focal	kincode	relation	kin	source (c_source → TEXT_CODES)
10884 陳俊卿	243	孫 (SS, son's son, ↓2)	15213 陳文龍	7596 宋人傳記資料索引（電子版）
15213 陳文龍	62	祖父 (FF, ↑2)	10884 陳俊卿	0 未知 (none)
15201 陳瓚	79	從父/伯叔父 (FB)	15213 陳文龍	2229 宋史
15213 陳文龍	177	從子/姪 (BS)	15201 陳瓚	2229 宋史

2. Source trace-back — every layer says "descendant/collateral," none says "grandson"

Layer	Wenlong ↔ Junqing	Term
宋史 卷451 陳文龍傳 (text behind source 7596)	「陳文龍字君賁，福州興化人。丞相俊卿之後也。」	之後 = descendant (generation-open)
Scholarship / 玉湖陳氏 genealogy	「為陳俊卿五世從孫」；父＝陳粢	五世從孫 = collateral
CBDB	kincode 243 (孫, ↓2, lineal)	孫 = grandson

3. The lineage in detail (玉湖陳氏 / Yuhu Chen)

The clan's two Southern Song chancellors — Junqing and Wenlong — are the source of the epithet 「一門二丞相，九代八太師」. They are not lineal grandfather/grandson; they head two branches that diverge at the 2nd-generation ancestor 貴 (Chen Gui). Yuhu-generation numbers; CBDB IDs and posthumous 太師 dukedoms shown where attested:

Gen 1  仁  Chen Ren (1015–1064)            — Yuhu founder; posth. 太師 沂國公
Gen 2  貴  Chen Gui                        — posth. 太師 蜀國公
        │
        ├── Branch A (詵 line) ───────────────────────────────────────────
        │   Gen 3  詵  Chen Shen (CBDB 10883)    — posth. 太師 冀國公
        │   Gen 4  俊卿 Chen Junqing (CBDB 10884, 1113–1186) ★ Chancellor #1; 太師 魏國公
        │   Gen 5  宓  Chen Mi (CBDB 10892, 1171–1230)   — Junqing's 4th son
        │   Gen 6  〔Mi's son〕
        │   Gen 7  瓚  Chen Zan (CBDB 15201, d. 1277) ★ martyr
        │          └─ CBDB FF = 宓 (10892): Zan is Junqing's great-grandson — chronologically clean
        │
        └── Branch B (德戴 line) ──────────────────────────────────────────
            Gen 3  德戴 Chen Dedai                — brother of 詵
            Gen 4  宋卿 Chen Songqing
            Gen 5  欽紹 Chen Qinshao             — posth. 太師 永國公
            Gen 6  家/衰/寬 Chen Jia             — posth. 太師 安國公
            Gen 7  粢  Chen Zi                   — posth. 太師 榮國公   [Wenlong's father]
            Gen 8  文龍 Chen Wenlong (CBDB 15213, 1232–1277) ★ Chancellor #2; 太師 忠肅

Two things this makes clear:

• The "八太師" is a cross-branch prestige roster, not a patriline. The eight 太師 titles run gens 1–4 on Branch A (仁→貴→詵→俊卿) and gens 5–8 on Branch B (欽紹→家→粢→文龍); six are posthumous ennoblements conferred when each chancellor was honored. The 沂/蜀/冀國公 (Branch A) and 永/安/榮國公 (Branch B) dukedoms independently corroborate the two-branch shape. Collapsing this roster into a literal 祖→孫 is, we believe, how the 孫 edge arose.
• Zan (15201) is the genuine lineal descendant; Wenlong is the collateral one. CBDB's own record places Zan's grandfather as 宓 (Junqing's son) — sound. Wenlong descends on Branch B, with no father currently recorded in CBDB (KIN_DATA has no F row for 15213); his father is 陳粢 (Chen Zi). The 文龍↔瓚 uncle/nephew tie (《宋史》-sourced) is therefore a cross-branch relation and is correct as-is.

(Counting note: the genealogy places Wenlong four generations below Junqing; scholarship's 「五世從孫」 uses inclusive 世-counting. Either way the relation is collateral, not 祖/孫.) This reconstruction is concordant across three editions in our corpus: the family manuscript copy (002-right), the 1937 Waihai Zupu Gao (外海鄉陳氏族譜稿, leaves 0007–0008), and the 1971 Utah Genealogical Society microfilm (frames 0012/0021, which states the collateral relation explicitly).

4. Why the lineal edge is doubtful (internal checks)

• No father for 15213 (no F row) — Wenlong is attached straight to Junqing with intervening generations absent.
• Provenance asymmetry: lineal edge cites only an index (7596) one way, 未知 the other; the collateral 文龍↔瓚 edge is《宋史》-sourced both ways.
• Chronology: Junqing b. 1113, Wenlong b. 1232 → 119 yrs for a claimed 2-generation link (~60 yrs/gen). CBDB also makes Wenlong (b. 1232) the uncle of Zan (c_index_year 1231) — an uncle the age of his nephew. Across the genealogy's branch spacing (~30 yrs/gen) the dates are consistent.

5. Suggested fix

• Re-map or remove the lineal edge between 15213 ↔ 10884 (kincodes 243 / 62); sources support a collateral descendant (之後 / 五世從孫), not 祖父/孫.
• Add Wenlong's father 陳粢 (Chen Zi) so the Branch B descent models correctly.
• Keep the 《宋史》-sourced 文龍↔瓚 uncle/nephew edge.
• Fix surface (matches your workflow): https://input.cbdb.fas.harvard.edu/basicinformation/15213/kinship

6. Reproduce

SELECT k.c_personid, bf.c_name_chn AS focal, k.c_kin_code, kc.c_kinrel_chn,
       k.c_kin_id, bk.c_name_chn AS kin, k.c_source, t.c_title_chn
FROM KIN_DATA k
JOIN KINSHIP_CODES kc ON kc.c_kincode=k.c_kin_code
LEFT JOIN BIOG_MAIN bf ON bf.c_personid=k.c_personid
LEFT JOIN BIOG_MAIN bk ON bk.c_personid=k.c_kin_id
LEFT JOIN TEXT_CODES t ON t.c_textid=k.c_source
WHERE k.c_personid IN (10884,15213,15201) ORDER BY k.c_personid;
-- no father for Wenlong →
SELECT * FROM KIN_DATA k JOIN KINSHIP_CODES kc ON kc.c_kincode=k.c_kin_code
WHERE k.c_personid=15213 AND kc.c_kinrel='F';  -- 0 rows

I have the multi-edition genealogy reconstruction (family copy + 1937 + 1971) showing the 貴→德戴→宋卿→欽紹→家→粢→文龍 line and can share scans/transcriptions on request.

— Reggie Chan (reggie.chan@gmail.com), Yuhu / Waihai Chen genealogy digitization project

Sources: 宋史卷451 (CTEXT chapter 658565); 陳文龍, zh.wikipedia; CBDB bulk SQLite (KIN_DATA/KINSHIP_CODES/BIOG_MAIN/TEXT_CODES).

[sudoghut]

Hi Reggie,

Thank you so much for raising this issue! This is a very interesting case. If a CBDB record is dated 2007, it suggests that the data may have originated from our project founder, the late Professor Robert Hartwell, or one of his collaborators, and therefore predates our work on the 宋人傳記資料索引.

After further research, I found that the relationship between them is recorded very differently across sources:

祖 FF(SS) - 昭忠錄
五世從孫 SSSSS - 八閩通志
從兄弟之玄孫 SSS - 閩書
高祖 FFFF(SSSS) - 蒲陽文獻

In the absence of strong evidence to the contrary, CBDB's policy is to follow its primary sources and record differing information from other sources in the c_notes field, allowing users to make their own informed decisions.

I also cited this issue in the c_notes field of the following record: https://input.cbdb.fas.harvard.edu/basicinformation/15213/kinship/edit?c_personid=15213&c_kin_id=10884&c_kin_code=62

Thanks!
Hongsu

[reggiechan74]

Thank you Hongsu. Please see attached images from 1937 Waihai Chan zupu, which I got from Harvard (your site). note the 4707,4708,4709 numbers you see below are my database numbers not yours. your CBDB numbers are referenced.

仁 #0 (gen1)
└ 貴 #1 (gen2)
├─ 詵 #3 (gen3, CBDB 10883)
│ └─ 俊卿 JUNQING #10 (gen4, CBDB 10884, b.1113) ← Prime Minister #1
└─ 德戴 #2 (gen3)
└─ 宋卿 #5 (gen4)
└─ 欽紹 #12 (gen5)
└─ 衮 #4707 (gen6)
└─ 粢 #4708 (gen7)
└─ 文龍 WENLONG #4709 (gen8, CBDB 15213, b.1232) ← Prime Minister #2

So the verdict on "grandson or something else":

• Junqing is gen 4; Wenlong is gen 8 — four generations apart, ~119 years (1113 → 1232).
• Their lines split at 貴 (gen 2): Junqing through 詵, Wenlong through 詵's brother 德戴. They are collateral,
  sharing a common ancestor two generations above Junqing.

Trace the lineage yourself using the cross-referencing done by given name/courtesy name or literary name. Path A leads to JQ and Path B leads to WL.

Feel free to add this analysis to your primary sources, but a solid tree showing the father-son chain visually is also difficult to argue with.
0022-annotated.jpg
0023-annotated.jpg