Determine string language or letters

You'll still need to do some additional filtering to check that there are some Latin letters in there (A to Z => 65 to 90 , a to z => 97 to 122). 

For example "%!$*?" is gibberish in any language but the characters all have unicode values less than 128.

following Alan's suggestion,  You could use NetSpell (a .NET open source library that checks spelling).  If the word doesn't appear here, then there is a good chance it's not English.  The problems will come when you try to identify a word that is not English, but still appears in the English dictionary.  It's helpful to have cues such as the font being used or the LCID (Local Identifier)

It's easy to detect whether a string contains at least one Latin letter with this regular expression:

using System.Text.RegularExpressions;

........

string s = "abcABC"; // or whatever
Regex r = new Regex("[a-zA-z]+");
bool containsLatinLetters = r.IsMatch(s);
       
It's not difficult to extend this technique to check that a string contains no other characters apart from Latin letters, numerals and punctuation used in the English language.

However, having done that, it would still be difficult to check that the string was actually written in English. A possible technique would be to parse the string into individual words and check whether they all occurred in an English dictionary or not.