How the “Sounds Like” Search Matches Names

CRARG records stretch from the late 1700s to the 1950s. In that time, records appear in Polish, Russian (using the Cyrillic alphabet), Polish again, German (in part of the 1910s and of course in the 1940s), in English (during and after the Holocaust era), and in Yiddish or Hebrew (using Hebrew letters).

Searching by exact spelling will very often miss records that belong to your family. Here are some examples:

• In the early 1800s, surnames beginning with “I” began with what looked like a “J,” so Ickowicz may seem to appear as Jckowicz.

• Cyrillic has no letter for “h,” so in Russian-language documents Hofman appears as Gofman or Chofman.

• Cyrillic has no exact equivalent for the Polish letter “ę,” so in Russian-language documents Okręt appears as Okrent.

• Polish records add grammatical endings, so surnames appear in forms such as Goldbergowa (wife of Goldberg) or Goldbergówna (daughter of Goldberg).

• If a family migrates to a new country, a common surname like Jakubowicz may be changed to Yakubovich or Jacobowitz.

With CRARG records, when you search by “sounds like” each name is converted to a phonetic form and compared by how it is pronounced, rather than how it is spelled. This lets you search and find a surname across the Polish, Russian, German, English, Yiddish, and Hebrew spellings that appear in the records, while still keeping genuinely different names apart.

The examples below are computed live by the search itself. Each pair shows its phonetic reading and a distance score; names at or below 1.0 are treated as the same name, higher scores as different names.

Same name, different spelling

Names are read by sound, not letters, so spellings from Polish, German, English, and Yiddish sources match each other. Polish diacritics and their stripped forms are the same name (ą = a, ś = s, ż = z), and stray accents from other languages are ignored (í = i, é = e).

Spelling	Spelling	Phonetic reading	Distance	Result
Jakubowicz	Yakubovitch	/jakubɔvit͡s/ · /jakubɔvit͡ʃ/	0.08	same name
Fiszman	Fischman	/fiʃman/ · /fiʃman/	0.0	same name
Szwarc	Schwarz	/ʃvart͡s/ · /ʃvart͡s/	0.0	same name
Kohn	Cohn	/kɔn/ · /kɔn/	0.0	same name
Katz	Kac	/kat͡s/ · /kat͡s/	0.0	same name
Altman	Ałtman	/altman/ · /altman/	0.0	same name
Altman	Aldtman	/altman/ · /altman/	0.0	same name
Dykerman	Dickerman	/dɨkɛrman/ · /dikɛrman/	0.1	same name
Jarząbek	Jarzabek	/jaʒɔ̃bɛk/ · /jaʒabɛk/	0.25	same name
Zając	Zajac	/zajɔ̃t͡s/ · /zajat͡s/	0.29	same name
Lewin	Lewín	/lɛvin/ · /lɛvin/	0.0	same name
Szwarcberg	Schwarzberg	/ʃvart͡sbɛrɡ/ · /ʃvart͡sbɛrɡ/	0.0	same name

A written i or j can be pure spelling

Polish and Russian spellings insert an i or j after a consonant to mark a softened sound; it comes and goes across sources (Pacianowski = Pacanowski, Paljuch = Paluch, Dyjament = Dyament).

Spelling	Spelling	Phonetic reading	Distance	Result
Pacanowski	Pacianowski	/pat͡sanɔvski/ · /pat͡sanɔvski/	0.0	same name
Paluch	Paljuch	/palux/ · /palux/	0.0	same name
Klug	Kliug	/kluɡ/ · /kluɡ/	0.0	same name
Landau	Ljandau	/landau/ · /landau/	0.0	same name
Dymant	Diament	/dɨmant/ · /damɛnt/	0.38	same name

Grammatical endings are ignored

Slavic genitive and feminine endings (-ów, -owa, -owej, -owie, -ówna, -ova, -aja) mean "of" or "daughter of", so they are matched as the base surname; a fleeting "e" returns when the ending comes off (Galstrow = Galster).

Spelling	Spelling	Phonetic reading	Distance	Result
Goldberg	Goldbergowa	/ɡɔldbɛrɡ/ · /ɡɔldbɛrɡ/	0.0	same name
Goldberg	Goldbergówna	/ɡɔldbɛrɡ/ · /ɡɔldbɛrɡ/	0.0	same name
Klug	Klugów	/kluɡ/ · /kluɡ/	0.0	same name
Abramowicz	Abramowitz	/abramɔvit͡s/ · /abramɔvit͡s/	0.0	same name
Moskowicz	Moskowiczowey	/mɔskɔvit͡s/ · /mɔskɔvit͡s/	0.0	same name
Galster	Galstrow	/ɡalstɛr/ · /ɡalstr/	0.57	same name

One -stein, many transliterations

The German -stein ending appears as -sztajn, -sztyn, -szten, -stien and more, with the vowel before it freely reduced.

Spelling	Spelling	Phonetic reading	Distance	Result
Epsztajn	Epsztyn	/ɛpʃtajn/ · /ɛpʃtajn/	0.0	same name
Epsztajn	Epszten	/ɛpʃtajn/ · /ɛpʃtajn/	0.0	same name
Borensztajn	Bornsztyn	/bɔrnʃtajn/ · /bɔrnʃtajn/	0.0	same name

German spelling conventions

An unstressed middle "e" is silent (Tillemann = Tilman, Auerbach = Aurbach), "ae" is the vowel ä, and between vowels f = ff = pf.

Spelling	Spelling	Phonetic reading	Distance	Result
Talman	Tillemann	/talman/ · /tilɛman/	0.88	same name
Talman	Teleman	/talman/ · /tɛlɛman/	0.81	same name
Telman	Taelman	/tɛlman/ · /tɛlman/	0.0	same name
Orbach	Auerbach	/ɔrbax/ · /aurbax/	0.46	same name
Fajerman	Faiermann	/fajrman/ · /fajrman/	0.0	same name
Kifer	Kupfer	/kifɛr/ · /kufɛr/	0.3	same name
Kifer	Kiffer	/kifɛr/ · /kifɛr/	0.0	same name

The digraph "ie" is read by language

In Polish names "ie" is a glide (as in Sieradzki); in German and Yiddish names it is a long i, so Wielingier matches Willinger.

Spelling	Spelling	Phonetic reading	Distance	Result
Willinger	Wielingier	/vilinɡɛr/ · /vilinɡir/	0.09	same name
Englender	Engielender	/ɛnɡlɛndɛr/ · /ɛnɡilɛndɛr/	0.2	same name

Silent lengthening "h"

In German and Ladino spellings an "h" after a vowel marks length and is silent, so Kohn is Kon and Cohen is Koen.

Spelling	Spelling	Phonetic reading	Distance	Result
Kohn	Kon	/kɔn/ · /kɔn/	0.0	same name
Cohen	Koen	/kɔhɛn/ · /kɔɛn/	0.8	same name

H, G, and Ch interchange

Slavic languages have no "h" sound, so an H name surfaces with G or Ch in Polish and Russian records. The bridge is the H itself: native G and Ch names stay separate.

Spelling	Spelling	Phonetic reading	Distance	Result
Horowicz	Górewicz	/hɔrɔvit͡s/ · /ɡurɔvit͡s/	0.75	same name
Hofman	Gaufman	/hɔfman/ · /ɡaufman/	0.94	same name
Hofman	Chaufman	/hɔfman/ · /xaufman/	0.94	same name

The diphthong "au" flattens to o

German and Yiddish "au" is spoken and spelled as plain o or u across sources; the -baum ending appears as -bom, -boim, -bojm, -bajm.

Spelling	Spelling	Phonetic reading	Distance	Result
Krause	Kroze	/krausɛ/ · /krɔzɛ/	0.5	same name
Landau	Lando	/landau/ · /landɔ/	0.46	same name
Hofman	Haufman	/hɔfman/ · /haufman/	0.69	same name
Birnbaum	Birenbom	/birnbaum/ · /birɛnbaum/	0.4	same name
Birnbaum	Birenbojm	/birnbaum/ · /birɛnbaum/	0.4	same name

Hebrew and Yiddish script

Names written in Hebrew letters are transliterated and matched against their Latin-letter spellings.

Spelling	Spelling	Phonetic reading	Distance	Result
Willinger	ווילינגער	/vilinɡɛr/ · /vilinɡɛr/	0.0	same name
Fiszman	פישמאן	/fiʃman/ · /fiʃman/	0.0	same name

Y and J are the same sound

A leading "Y" before a vowel is the same consonant as "J".

Spelling	Spelling	Phonetic reading	Distance	Result
Yankel	Jankel	/jankɛl/ · /jankɛl/	0.0	same name
Yosef	Josef	/jɔsɛf/ · /jɔsɛf/	0.0	same name
Yurkowicz	Jurkiewicz	/jurkɔvit͡s/ · /jurkɔvit͡s/	0.0	same name

Different names stay separate

Names that merely share a Soundex code but genuinely sound different are kept apart, so results stay precise. The first and last consonants carry a name's identity, and two independent sound changes stack into a different name even when each alone would be a plausible variant (Kupfer is Kifer, but Kuper — vowel shift AND p for f — is not).

Spelling	Spelling	Phonetic reading	Distance	Result
Zajdman	Zytman	/zajdman/ · /zɨtman/	2.13	different
Kac	Koch	/kat͡s/ · /kɔx/	3.22	different
Szwarc	Swierk	/ʃvart͡s/ · /svirk/	2.94	different
Brat	Broda	/brat/ · /brɔda/	1.1	different
Besser	Fiszer	/bɛsɛr/ · /fiʃɛr/	1.6	different
Berkowicz	Perkowicz	/bɛrkɔvit͡s/ · /pɛrkɔvit͡s/	1.06	different
Kon	Gan	/kɔn/ · /ɡan/	1.5	different
Borensztajn	Bronsztajn	/bɔrnʃtajn/ · /brɔnʃtajn/	3.38	different
Gitler	Chitler	/ɡitlɛr/ · /xitlɛr/	1.21	different
Gutman	Chutman	/ɡutman/ · /xutman/	1.42	different
Kifer	Kuper	/kifɛr/ · /kupɛr/	1.05	different
Strauss	Sztrauch	/straus/ · /ʃtraux/	1.02	different

Beyond the score: rarity and misspellings

A few kinds of spellings are judged by more than their sound score. A spelling that just misses the cutoff is still included when it accounts for only a tiny share of the results — a near-miss that adds one record costs nothing to show, while the same score carrying hundreds of records would swamp the results and stays out. Obvious scribal slips — two letters swapped (Szymkoiwcz), or a dropped vowel that leaves an unpronounceable letter pile (Ickwicz) — are included when the spelling is rare, since a handful of records with an impossible spelling can only be a copying error, whereas a frequent spelling is a real name in its own right. And some differences are never bridged by rarity: a name's first and last consonants carry its identity, and an H- name may surface as G- or Ch- in Russian-era records but never the reverse.

Matching runs on top of a Daitch-Mokotoff Soundex refined over two decades of use; the phonetic reading shown here is what further separates real spelling variants from names that only happen to share a Soundex code.