How to draw a kinship diagram, step by step

A kinship diagram is a map of who is related to whom, drawn with a handful of shapes and lines. In one glance it shows what would take a page of prose: which cousins share a grandfather, which marriage produced which child, who has died, and where the person you are talking to sits in all of it.

Anthropologists have drawn these charts since systematic fieldwork began. W. H. R. Rivers worked out his genealogical method during the Cambridge expedition to the Torres Strait in 1898 and described it in print in 1910, and collecting a genealogy and drawing it is still one of the first practical skills taught in kinship courses.

The good news for a beginner is that the symbol set is small. The less good news is that a few conventions differ between textbooks and between disciplines. So the habit that matters most is a dull one: every diagram gets a key.

The symbol set

Most anthropology texts use the conventions below. Where two forms are common, the table gives both.

If you have seen a family tree in a genetics clinic or a biology textbook, you have met a close relative of this system with one difference that trips people up. Clinical pedigrees follow recommendations that the National Society of Genetic Counselors first published in 1995 and has updated since. In that system a square is a male, a circle is a female and a diamond is a person whose sex is not specified. A diagonal line through a symbol marks a death, an arrow points to the person who brought the family to the clinic, siblings run from oldest on the left to youngest on the right, and a break in the line between two partners shows that the relationship has ended.

So a square on an anthropologist's chart and a square on a genetic counselor's chart can stand for different people. Neither convention is wrong. Mixing them without a key is.

Kin-type strings: a name for every position

Diagrams and kin-type notation work as a pair. Each relative gets a string of letters that spells out the path from Ego: F father, M mother, B brother, Z sister, S son, D daughter, H husband, W wife, with e for elder and y for younger. Read the string from left to right, starting at Ego. FZD is father's sister's daughter. MBW is mother's brother's wife. FeB is father's elder brother, which says far more than "uncle", an English word that also covers MB, FZH and MZH.

That precision is the reason to bother. Kin terms differ from language to language, and many terminologies group relatives in ways English does not; the six classic terminology types show how differently the same family can be carved up. If you label positions with kin types first, you can then record what Ego actually calls each person without your own language's categories sneaking into the data. Our notation guide has the full set of letters, along with conventions such as ms and ws (man speaking, woman speaking).

Worked example: from an interview to a diagram

Below is the kind of account you might write down during a genealogical interview. It is entirely hypothetical. The speaker is a married woman, and we will call her Ego.

My father is the youngest of three. His sister is the eldest. She married young and divorced, and she has a daughter from that first marriage. Later she remarried and had a son with her second husband. Next comes my father's brother, who died six years ago. His widow still lives in the family house, and they had two sons. My mother has one brother, younger than her, who never married. I have an older brother. Between him and me my mother had a baby who died a few days after birth, and I was never told whether it was a boy or a girl. I'm married, and we have a daughter and, after her, a son.

That is eighteen people across three generations. A clean drawing comes from doing the work in a fixed order rather than starting at the top left and hoping.

Step 1: list everyone and give each a kin type

Two things already stand out. Two different men share the kin type FeZH, so give them working labels. And the baby has no single-letter code, because B and Z both assume a known sex.

Step 2: give each generation its own row

Draw three faint guide lines across the page. Ego's parents' generation goes on the top line (G+1), Ego's own generation in the middle (G0), and Ego's children at the bottom (G-1). People who married in belong on the row of the person they married, whatever their age.

Step 3: place the couples before anyone else

Start with Ego's parents near the middle of the top row: a triangle and a circle joined by a double line. Then work outward. The father's siblings go on his side and the mother's sibling on hers, so that no line will ever have to cross the marriage between F and M. Here the account helps: because the father is the youngest of three, his elder sister and elder brother can sit to his left in birth order, eldest furthest out. The mother's younger brother goes to her right.

FeZ has two marriages. Put H1 on one side of her and H2 on the other, and mark the divorce on the first marriage line. Each marriage gets its own descent line, which makes it plain that FeZD and FeZS are half siblings: same mother, different fathers.

Step 4: hang the children

From the middle of each marriage line, drop a vertical line to a sibling bar and hang the children from it, oldest on the left where you know the order. Ego's own sibling set runs eB, then the baby, then Ego. Her husband sits beside her on the middle row, and their daughter and son hang from that marriage on the bottom row.

Step 5: mark what is not about position

Now add the status marks: the diagonal line through FeB and through the baby, the divorce mark, the diamond for the baby whose sex is unknown, and Ego's filled symbol.

Step 6: read it back

Finally, test the drawing against the words. Point to each symbol, say its kin type aloud, and check that walking along the lines from Ego produces that string. If the path to someone you labeled FeBS runs through the wrong marriage, you will hear the mistake.

Here is the finished diagram as a plain-text sketch. To keep it typeable, /_\ stands for a triangle, ( ) for a circle and < > for a diamond, an x inside a symbol stands for the diagonal line of a death, (@) is Ego, ===== is a marriage and ==/== a marriage that ended in divorce.

           +-----------------+-----------------+         +-------+
           |                 |                 |         |       |
/_\ ==/== ( ) ===== /_\     /x\ ===== ( )     /_\ ===== ( )     /_\
H1    |   FeZ   |   H2      FeB   |  FeBW      F    |    M      MyB
      |         |                 |                 |
      |         |             +---+---+     +-------+-------+
      |         |             |       |     |       |       |
     ( )       /_\           /_\     /_\   /_\     <x>     (@) ===== /_\
    FeZD      FeZS          FeBS    FeBS   eB      sib     Ego   |    H
                                                                 |
                                                             +---+---+
                                                             |       |
                                                            ( )     /_\
                                                             D       S

Notice what the picture shows that the paragraph did not. The cousins come in two kinds. The two FeBS are linked to Ego through a pair of brothers, while FeZD and FeZS are linked through a brother and a sister, and many terminologies name those two kinds of cousin quite differently. The kinship terms chart lets you see how several systems would label these same positions, and the cousin calculator will count the degree and removal between any two people once you know their common ancestors.

Mistakes that show up again and again

• Hanging a child from one parent instead of from the marriage line. A child drawn under the mother alone reads as the child of a union you did not record.
• Letting lines cross when a different arrangement would avoid it. Most crossings disappear if you move a sibling set to the outer side of a couple. In large genealogies with marriages between relatives some crossings cannot be avoided; note them rather than bending the generation rows.
• Putting people on the wrong row because of age. A second wife twenty years younger than her husband still sits on his row.
• Leaving out the dead and the divorced. A brother who died before Ego was born may still matter for inheritance or naming, and a former marriage is the only thing that explains a half sibling.
• Labeling with English words instead of kin types. "Cousin" hides at least eight kin types (FBS, FBD, FZS, FZD, MBS, MBD, MZS, MZD), and the people you are studying may treat each of them differently.
• Mixing the triangle convention with the clinical square convention, with no key to say which is which.
• Writing real names on a diagram you will share or publish. Kin types and numbers protect the people who talked to you. Keep the list of names separate and secure.

Software, in general terms

A pencil and squared paper remain the quickest tools during an interview, because you can redraw as the person remembers more. For a clean copy, the options fall into roughly four groups:

1. General drawing and diagramming programs, where you place every shape and connector yourself. They are flexible, but nothing checks that the family you drew makes sense.
2. Genealogy programs, which store people and links as data and draw charts from them. Most can read and write GEDCOM, the long-standing exchange format for family data, now maintained by FamilySearch and currently at version 7.0.
3. Clinical pedigree tools, which follow the genetic counselors' symbols (squares, not triangles) and suit family health histories.
4. Research tools for kinship network analysis, which treat a genealogy as a graph and can search it for patterns such as marriages between relatives.

Whatever you choose, ask one question before you commit: can you get your data out in an open format? The drawing is a picture. The list of people and the links between them is the thing worth keeping.

A good first exercise is to draw your own family back to Ego's grandparents using kin types only, then ask a relative to read the chart back to you and catch what you missed.

Sources and further reading

• Rivers, W. H. R. (1910). The genealogical method of anthropological inquiry. The Sociological Review 3(1): 1-12. doi.org/10.1111/j.1467-954X.1910.tb02078.x
• Bennett, R. L., Steinhaus, K. A., Uhrich, S. B., O'Sullivan, C. K., Resta, R. G., Lochner-Doyle, D., Markel, D. S., Vincent, V., and Hamanishi, J. (1995). Recommendations for standardized human pedigree nomenclature. Journal of Genetic Counseling 4(4): 267-279; also printed in American Journal of Human Genetics 56: 745-752. doi.org/10.1007/BF01408073
• Bennett, R. L., French, K. S., Resta, R. G., and Doyle, D. L. (2008). Standardized human pedigree nomenclature: update and assessment of the recommendations of the National Society of Genetic Counselors. Journal of Genetic Counseling 17(5): 424-433. doi.org/10.1007/s10897-008-9169-9
• Saneda, T., and Field, M. Cultural Anthropology, section 2.3, Kinship. LibreTexts, CC BY-SA 4.0. socialsci.libretexts.org
• FamilySearch. GEDCOM specification, version 7.0. gedcom.io