There is such a small statistics

“n” people ask now help!

if (n == 1) = & gt; "1 person asks for now help"
elseif ($ n == 2) = & gt; "2 people are asking for help now"

and so on.

How to do it?

1, Authority 100%

function declofnum ($ num, $ titles) {
  $ Cases = Array (2, 0, 1, 1, 1, 2);
  Return $ num. "". $ Titles [($ num% 100 & gt; 4 & amp; & amp; $ num% 100 & lt; 20)? 2: $ Cases [min ($ num% 10, 5)]];
}
Echo Declofnum (5, Array ('man asks', 'a person is asked', 'man ask'));

2, Authority 35%

Here people update the rules for different languages: Language Pral Rules .

For Russian Language:

one → n mod 10 is 1 and n mod 100 is not 11;
FEW → N MOD 10 in 2..4 and n mod 100 not in 12..14;
Many → N MOD 10 IS 0 or N mod 10 in 5..9 or n mod 100 in 11..14;
Other → Everything ELSE

In this way:

function pluralcategory ($ count)
{
 $ MOD10 = $ Count% 10;
 $ mod100 = $ Count% 100;
 If (IS_INT ($ Count) & amp; & amp; $ MOD10 == 1 & amp; & amp; $ mod100! = 11) {
  Return 'One';
 } elseif (($ MOD10 & GT; 1 & amp; & amp; $ MOD10 & LT; 5) & amp; & amp; ($ MOD100 & LT; 12 || $ MOD100 & GT; 14)) {
  Return 'FEW';
 } elseif ($ MOD10 == 0 || ($ MOD10 & GT; 4 & amp; & amp; $ MOD10 & LT; 10) || ($ MOD100 & GT; 10 & amp; & amp; $ mod100 & lt; 15)) {
  Return 'Many';
 } else {
  Return 'Other';
 }
}

by itself, only the category will be in response. For the word “man” array of correspondences will be:

one → man
FEW → man
Many → Man
Other → Human

Demo:

$ WORD = Array (
 'one' = & gt; 'human',
 'FEW' = & GT; 'man',
 'Many' = & gt; 'human',
 'other' = & gt; 'man',
);
for ($ i = 1; $ i & lt; 500; $ i ++) {
 Echo $ i. ' '. $ Word [PluralCategory ($ i)].' & lt; br / & gt; ';
}

3, Authority 26%

Option for this case is excessable, but it is suitable for other cases when the word is not two, but three forms:

for ($ n = 1; $ n & lt; 4096; $ n ++)
{
  $ plural =.
  (
    $ n% 10 == 1 & amp; & amp; $ n% 100! = 11
    .
    'human'
    :
    (
      $ n% 10 & gt; = 2 & amp; & amp; $ n% 10 & lt; = 4 & amp; & amp;
      (
        $ n% 100 & lt; 10 || $ n% 100 & gt; = 20
      )
      .
      'man'
      :
      'human'
    )
  );
  Print "$ n $ plural \ n";
}

4, Authority 23%

Remarking too. Crossed the algorithm @Lampa with your templates. It seems to me so more convenient. Code for many lines due to a free recording of a regular expression, so just give a reference to the implementation with tests:
http://ideone.com/xqgk2g
The Sklonit function you need to simply pass the text containing how much templates:

// template 1
[Number: Start (end1, end2, end3)] // Start + end
[2: Apples (Oh, A,)] // Apples
// Template 2.
Number [Separator: Start (end1, end2, end3)] // Number + Separator + Start + End
2 [: apples (oh, a,)] // 2 apples

At the same time, if the separator in the second template is a number, then it is considered that this is the template of the first type, that is,

1 [2: apples (o, a,)] // 1Block, and not 12Ablok, if you perceive it like Template 2.