I changed your program as follows:

use strict;
use WWW::Mechanize;
use HTML::TokeParser;
use Data::Dumper;

my $webcrawler = WWW::Mechanize->new();
$webcrawler->get("http://www.marsh.com");
my $content = $webcrawler->content;
my $parser = HTML::TokeParser->new(\$content);
my @text_data;
while($parser->get_tag){
    push @text_data, $parser->get_trimmed_text(),"\n";
}
print Dumper(\@text_data);
[download]

And it works as advertized.

BTW, You better not crawl/scrape Google as they don't like it very much: see "No automated querying" (external link).