To Err is Human, to Encrypt Divine

In Minnesota, the number on your driver’s license is a function of your name and date of birth. A woman with the maiden name Linda Louise Eastman who marries James Paul McCartney and takes the name Linda Eastman McCartney sees her Minnesota driver’s license number change from E-235-522-549-898 to M-263-522-162-898 because her last and middle names change. Exactly how the name and number correspond is a closely guarded secret of the Minnesota Department of Public Safety. However, Minnesota and Michigan code their numbers using the identical method—and Michigan doesn’t keep the method secret. The first character of a person’s Minnesota driver’s license number is the initial of the surname (E for Eastman). The next three characters of the license number are obtained by applying a complicated system called the Soundex code.

Here’s how Soundex works: Delete all Hs and Ws, and assign numbers 0 through 6 to the remaining letters as follows: A, E, I, O, U, Y=0; B, F, P, V=1; C, G, J, K, Q, S, X, Z=2; D, T=3; L=4; M, N=5; R=6. Convert the letters to numbers. If two or more adjacent digits are the same, omit all but the first. Next, unless the first letter was an H or W, delete the first digit. Then, delete all 0s. Retain only the first three digits among the remaining digits. If you have fewer than three digits left, add zeros. For example, for the surname McCartney we have the following numbers at successive stages of the process: 522063500; 5206350; 206350; 2635; 263.

The next two blocks of digits are determined from tables by using parts of the first and middle names. In most cases the code is determined by the first two characters such as Aa (028) for Aaron or Ji (414) for Jill. Exceptionally common names such as Joseph or Robert have their own special numbers (441 and 745). Finally, the last three digits are determined from tables using the day and month of birth. The numbers usually alternate in increments of 2 and 3. For example, the codes for March 1, 2 and 3 are 159, 162 and 164, respectively.

What is the advantage of this system? It’s an error-correcting scheme. It’s designed so that likely misspellings of a name nevertheless result in the correct coding of the name. For example, frequent misspellings of Erickson such as Ericksen or Ericsen yield the correct code for Erickson (E-625). Likewise, the code for Kristen and Kristin is the same (478) and the code for Emilie and Emily is the same (229). So, if a police officer wanted to know the license number for Kristin Emilie Erickson but entered Kristen Emily Ericsen in the computer, the correct number would come up. The Soundex code was developed by the U.S. Census Bureau back when they still conducted much of their business in verbal interviews, and didn’t apparently have the time to check their spelling.