Linguistic pattern matching (LPM) searches represent an entirely new capability for working with large databases of scanned discovery documents or e-discovery.  Especially when combined with auto-coding, LPM represents the fastest, most precise way to begin analysis of large discovery databases.  Since the processing is completely automated, the databases can be ready to use within days of productions.

How Searches Are Performed 
With LPM, users select blocks of text to use as the basis for the search.  This can be done by highlighting text in litigation support software, transcript review software or in briefs being examined in Word.  LPM software analyzes the linguistic patterns contained in the selected text and finds the pages in the database containing patterns that most closely match those patterns.  Searchers do not need to use any specific search syntax such as Boolean operators, proximity operators, relational operators or wildcards.

How Results Are Presented
LPM results are ranked according to how closely the linguistic patterns in the documents match the source or comparison text.  Unlike traditional Boolean searches where the most relevant documents can literally be scattered at random throughout the results, with LPM the documents that are most like the comparison text are presented first, enabling searchers to focus their attention where it is most likely to yield the best results.

Does LPM Replace Current Litigation Support Software?
LPM is most often used as a supplement to existing litigation support software and in some cases is quite closely integrated with it.

Typical LPM Searches
As lawyers and paralegals become familiar with LPM, they find countless ways to use it.  Some standard benefits include:

Checking consistent redaction/privilege waiver.  LPM is an ideal way to check scanned documents and e-discovery to be sure that redaction has been done consistently.  With database creation being completely automated, producing parties can use this tool to perform redactions before the production is complete.  Requesting parties can work from redaction or privilege logs to quickly pull up redacted documents and check against other scanned documents or e-discovery to see if un-redacted versions have been produced. 

Easy search formulation and rapid results evaluation make this a far superior tool than Boolean searching, which can require substantial expertise and trial-and-error to refine searches — with redactions, there are simply too many things to check to be able to frame multiple Boolean searches.

Identifying “near” dupes in e-mail chains or clusters.  LPM lets lawyers and paralegals rapidly identify e-mail messages that were sent as a result of replying to, forwarding, or blind copying original e-mails.  It will also identify situations where the contents of messages have been copied and pasted into other e-mail messages or other documents.

Finding “lost” documents.  LPM is an ideal tool for identifying documents that a lawyer just “knows” he/she has seen before and for finding documents that are described by a witness.

About our author . . .

Joseph Howie is Director of Client Services for Syngence, LLC, of Dallas, Texas, provider of Synthetix™ linguistic pattern matching searching capability and Syndex™ auto-coding.  You may reach him at jhowie@syngence.com.