Soundex

From FamilySearch Wiki

(Difference between revisions)
(word smithing)
(genealogical)
Line 7: Line 7:
 
The indexing system was developed by Robert C. Russell and Margaret K. Odell. It was patented in 1918 by Russell, and again in 1922 by Russell and Odell. It is formally called the Russell Soundex, and a variation used on the censuses is called the American Soundex.<sup>1</sup> When a computer was used to generate a Soundex index card for a census it was called a Miracode, and the information listed was slightly different from handwritten Soundex cards for the same census.  
 
The indexing system was developed by Robert C. Russell and Margaret K. Odell. It was patented in 1918 by Russell, and again in 1922 by Russell and Odell. It is formally called the Russell Soundex, and a variation used on the censuses is called the American Soundex.<sup>1</sup> When a computer was used to generate a Soundex index card for a census it was called a Miracode, and the information listed was slightly different from handwritten Soundex cards for the same census.  
  
The most well-known use of Soundex is on parts of the 1880, 1900, 1910, 1920, and 1930 United States federal censuses. It is also used by the federal government for selected ship passenger arrival lists, certain Canadian border crossings, and some naturalization records. A few county governments have also used a version of Soundex for courthouse kinds of records. More recently, Ancestry.com and other Internet companies have featured a Soundex search for their huge online genealogical databases.  
+
The most well-known genealogical use of Soundex is on parts of the 1880, 1900, 1910, 1920, and 1930 United States federal censuses. It is also used by the federal government for selected ship passenger arrival lists, certain Canadian border crossings, and some naturalization records. A few county governments have also used a version of Soundex for courthouse kinds of records. More recently, Ancestry.com and other Internet companies have featured a Soundex search for their huge online genealogical databases.  
  
 
Many non-genealogical search engine algorithms borrow heavily from concepts first introduced by Soundex.<sup>2</sup>
 
Many non-genealogical search engine algorithms borrow heavily from concepts first introduced by Soundex.<sup>2</sup>

Revision as of 06:18, 31 July 2008

The Russell Soundex (a.k.a. American Soundex, and Miracode) and its usefulness to genealogists are explained, some online Soundex converters listed, and rules given for how to manually create a Soundex code.

Contents

Definition and Value

Soundex is a phonetic index that groups together names that sound alike but are spelled differently, for example, Stewart and Stuart. This helps searchers find names that are spelled differently than originally expected, a relatively common genealogical research problem.

The indexing system was developed by Robert C. Russell and Margaret K. Odell. It was patented in 1918 by Russell, and again in 1922 by Russell and Odell. It is formally called the Russell Soundex, and a variation used on the censuses is called the American Soundex.1 When a computer was used to generate a Soundex index card for a census it was called a Miracode, and the information listed was slightly different from handwritten Soundex cards for the same census.

The most well-known genealogical use of Soundex is on parts of the 1880, 1900, 1910, 1920, and 1930 United States federal censuses. It is also used by the federal government for selected ship passenger arrival lists, certain Canadian border crossings, and some naturalization records. A few county governments have also used a version of Soundex for courthouse kinds of records. More recently, Ancestry.com and other Internet companies have featured a Soundex search for their huge online genealogical databases.

Many non-genealogical search engine algorithms borrow heavily from concepts first introduced by Soundex.2

Online Soundex Converters

The easiest way to obtain the Soundex code for a name is to use one of several online Soundex converter programs. Simply type a name, and at the click of a button, the converter will divulge the corresponding Soundex code. There may be subtle differences between programs:

How Soundex Works

Soundex is based on the classification of letters of the alphabet (consonants) into six sound-alike key letter groups. For example, in many languages the B and V sounds are nearly interchangeable; as are B and P; and V and F. So the first phonetic group of key letter consonants is b, f, p, v. Vowels are fluid and disregarded, as are H and W. By giving the same value to key letter consonants that often sound alike, the index brings names together that would usually be pronounced alike with little regard to their actual spelling. Each sound-alike group of key letter consonants is assigned a number. Each family name is assigned a Soundex code that has the initial letter of the name followed by exactly three of the sound-alike key letter group numbers. For example, Stewart = S363 and Stuart = S363.

Modern online search engines that use Soundex do it without displaying the Soundex codes—similar names spelled differently simply appear together on the search results list. But from time to time a researcher may need to understand Soundex codes in order to use one of the older Soundex indexes on microfilm.

Soundex Rules

Use these rules to manually create a Soundex code for an ancestor’s name.

Basic Rules3

  • Every soundex code consists of a letter and three numbers, such as D432.
  • The letter is always the first letter of the name. For example, Clausen = C425, and Klausen = K425.
  • After the first letter, disregard vowels (a, e, i, o, u, and y) and ignore the consonants h, and w.
  • Numbers are assigned to the remaining letters of the name according to the table of Soundex Key Letter Codes shown below.
  • Zeroes are added at the end if necessary to produce a four-character code. Excess letters are disregarded if they would produce a code longer than four-characters. For example Lee = L000, and Christopherson = C623.
Soundex Key Letter Codes
Number
    Represents the Letters
1
  b, f, p, v
2
c, g, j, k, q, s, x, z
3
d, t
4
l
5
m, n
6
r

Additonal Rules

  • Double key letters should be treated as one letter. For example, Gutierrez = G362.
  • Side-by-side letters with the same code number should be treated as one letter. For example, Campbell = C514, and Jackson = J250, and Pfister = P236.
  • Vowel key letter seperators. If a vowel (a, e, i, o, u, y) separates key letters that have the same code number, those key letters should be treated as two letters. For example, Tomzak = T522, and Roses = R220.
  • H or W key letter seperators. If an h or w separates key letters that have the same code number, those key letters should be treated as one letter. For example, Ashcroft = A261, and Carwruth = C630.
  • Names with prefixes, such as Van, Con, De, Di, La, or Le, are coded both with and without the prefix because the name might be listed under either code. Note, however, that Mc and Mac are not considered prefixes. For example, Van Deusen = V532 or D250.

More Soundex Examples4

Name Code       Name Code       Name Code       Name Code
Allricht A462 Kavanagh K152 McGee M200 Schafer S160
Eberhard E166 Lind L530 O'Brien O165 Shaeffer S160
Hanselmann H524 Lukaschowsky L222 Oppenheimer O155 Zita Z300
Heimbach H512 McDonnell M235 Riedemanas R355 Zitzmeinn Z325

Limitations and How to Overcome Them

Indexing rules were not always followed consistently. The government indexers may have occasionally overlooked some of the fine points of the additional indexing rules.5 If you cannot find a name under the correct code, try looking under the code as if the additional rules were overlooked. For example, try looking for Ashcroft under both A226 and A261, or try looking for Pfister under both P236 and P123.

Unrelated names may be grouped together. Sometimes names that do not appear to be related show up together on a Soundex index. Ignore clearly unrelated names. For example, if you were looking for Wilkins, you may also find under the same Soundex code, W425, the name Walakynowski.

Related names may not be grouped together. Sometimes names that are obviously related do not come together in the same Soundex index group. For example, Clausen is under C425 and Klausen under K425. If you cannot find a name you seek in a Soundex index, there are 20 alternative ideas in the Wiki article Guessing a Name Variation to help find elusive names in indexes.

Census Soundex Cards Show Limited Data

One of the most well-known uses of Soundex indexes is for some of the federal censuses of the United States. More recently, these old microfilm indexes have been largely replaced by online search engines. If you ever have an occasion to use a census Soundex on microfilm, keep in mind that the Soundex card is only a summary. It does not show as much information as the original census schedule.

Related Content

Endnotes

        1. Rick Parsons, Soundex - the True Story, (http://west-penwith.org.uk/misc/soundex.htm&nbsp;: accessed 30 July 2008).

        2. Ibid.

        3. Based on rules in The Soundex Indexing System, The National Archives (http://www.archives.gov/genealogy/census/soundex.html&nbsp;: accessed 30 July 2008).

        4. Anne Bruner Eales, and Robert M. Kvasnicka, Guide to Genealogical Research in the National Archives, 3rd ed. (Washington, D.C.: NARA, 2000), 22.

        5. Parsons.