2D arrays can be done in lpc, quite simply.  Just treat them as an array
of arrays.  So,
<pre>
   a = allocate(10);
   a[0] = allocate(10);
</pre>
etc...
<p>
and you can reference array 0, element 0, by
<pre>
   a[0][0]
</pre>
That's been around since arrays have been around.  You can't declare an 
array of more than 1 dimension (using the type * notation, if you have
type checking on), but you can have array of more than one dimension.  If
you have type checking on, you have to declare them as mixed * type, though.
<p>
This also works:
<pre>
mixed *a;

a = ({ ({1, 2, 3}), ({1, 2, 3}) });
</pre>
So, a[0] would be ({1, 2, 3}), and a[0][2] would be 3.
<p>
Or this:
<pre>
mixed *a;  /* please note the mixed * declaration is only for type checking 
              and means that this is an array of any type */

a = ({0, 0, 0, 0});  /* just to get the array to size 4 */
a[0] = ({1, 2, 3});
a[1] = ({1, 2, 3});
etc...
</pre>
I use multidimensional arrays everywhere, usually doing things in a loop:
<pre>
a = allocate(10);
for (i=0; i<10; ++i)
{
   string str;
   str = read_file(path, i+1);
   str = extract(str, 0, strlen(str)-2); /* to get rid of the \n */
   a[i] = explode(str, '#');
}
</pre>

(thanks to John Price, a.k.a. Raistlin for this explanation)