To provide a name extractor which can automatically extract the names including the person and commodity names, etc., consisting of the sounds accordant with each desired image.

A name extractor 1 includes a setting screen 2 where one or ≥2 concepts showing the images that can be sensed from names are designated, a 1st storage means 6 which stores the correspondence to the standard word sound elements showing the images included in those concepts and name sounds, a 2nd storage means 7 which stores plural names and a name extracting means 8 which selects the word sound elements necessary for a designated concept out of the means 6, compares the value of word sound elements selected among all names stored in means 7 with the standard value to calculate the degrees of satisfaction of the designated concepts and then arranges and outputs the names according to these degrees of satisfaction.