When I have encountered this problem in practice, there is usually a requirement to also find words that may or may not have trailing commas or hyphens (and sometimes where a comma doesn’t have trailing white space, etc.). I prefer to pre-process the source data to simplify the query and avoid having to enumerate all the possible edge cases.

In a generic environment with the requirement to find “words” that may be surrounded by any combination of space, hyphen, period, and/or comma, the simplest solution I have found is to replace all the “non-text” stuff with a separator character that isn’t in the desired text set and then use that as a start/end symbol when searching the text.

<searchcol> = ‘|’ + replace(replace(replace(replace(<sourcestring>,’,',’|'),’-',’|'),’.',’|'),”,’|') +’|’

Then searching with “… where <searchcol> like ‘%|’+ <word> + ‘|%’ …” will find that word, regardless of what non-text characters may precede or follow the word in the original source string.

The only way to find “words” is to create a second column containing the “words” separated (and preceded/followed!) by a known token, e.g., “|”.

You can do this with a trigger – when records are inserted into the table, generate the separate column with:
<newcol> = ‘|’ + replace(<name>, ‘ ‘, ‘|’) + ‘|’

Then, you can find the “word” of interest using:
select .. where <wordscol> like ‘%|lean|%’