I used the following lexicon:

d the
a quick
a brown
n fox
v jumps
p over
d the
a lazy
n dog
[download]

And the following input:

the quick brown fox jumps over the lazy dog
[download]

Running the following script with these two files as arguments

#!/usr/bin/perl
use warnings;
use strict;
use feature qw{ say };

my ($lexicon, $input) = @ARGV;

my %emotive;
open my $lex, '<', $lexicon or die $!;
while (<$lex>) {
    my ($pos, $lemma) = split;
    warn "Duplicate POS for $lemma.\n"
        if exists $emotive{$lemma} && $emotive{$lemma} ne $pos;
    $emotive{$lemma} = $pos;
}

my %seen;
open my $in, '<', $input or die $!;
while (<$in>) {
    $seen{$_}++ for map $emotive{$_}, split;
}

for my $pos (sort { $seen{$a} <=> $seen{$b} } keys %seen) {
    say "$pos\t$seen{$pos}";
}
[download]

the output is

p    1
v    1
n    2
d    2
a    3
[download]