In the spelling of English words, it is not uncommon to come across letters that are not sounded (or at least not directly) when we speak those words. Think the <h> in ‘hour’ and ‘honour’, the <k> in ‘knight’ and ‘knife’, and the <e> in ‘hate’ and ‘mate’. It is no wonder speakers of English are accused occasionally of being the worst spellers. Despite their apparent ‘silence’ in the speech of present-day speakers, unsounded letters speak volumes. They are windows to the ancestry of the words where they reside, influencers of the choice between grammatical articles ‘a’ and ‘an’, and signposts to pronunciation differences. This article showcases the depths of some of these silent letters.
To differentiate between letters and sound symbols, letters are kept in angle brackets <> while symbols representing sounds as they are spoken are kept in slashes //.
The silent <k> in knights of the round table
If we were to draw an analogy from human anatomy to talk about the letter <k> in English words like ‘knight’, ‘knife’, and ‘knee’, our best bet might be the appendix positioned in our lower right abdomen. Notwithstanding recent views in the medical field on its possible function, a long-held belief is that the appendix is a non-vital organ, “a useless remnant from our evolutionary past” (according to this article). Like the appendix, the letter <k> in ‘knight’, ‘knife’ and ‘knee’ is a remnant of linguistic evolution. Far from being useless, however, its presence in script is evidence that it was once pronounced, and it was for ease of articulation that people subsequently stopped saying the <k> in English words beginning with <kn>.
Indeed, this <k> was known to be sounded right through from the Old English period (around 476 AD–1066) to the beginning of the Early Modern English period (around 1476–1700, according to Campbell, 2004), although it would be unsurprising if the silencing of <k> began sometime in-between. In fact, the <k> in word-initial <kn> was probably no longer consistently heard by 1101 because by that time, the Danish name Knut was written Canut. The <ca> script basically shows that a vowel, i.e. the sound of <a>, had been inserted after the sound of <k>. A <ca> replacement of <k> word-initially before <n> is therefore also a sign that <k> in word-initial <kn> was once pronounced, so that if some semblance of it was still sounded, we would see the sound written as <ca> rather than just <k>.
Despite its silence, <k> in <kn>-initial English words remains vital in helping us identify the Germanic (as opposed to say, Latinate) ancestry of words such as ‘knight’, ‘knife’, and ‘knee’. This is due to a known Germanic influence on the English language coupled with the fact that <k> is pronounced still in present-day German words with an initial <kn>, e.g. in Knecht (German word meaning ‘servant’, which was what ‘knight’ at first referred to when it appeared in Old English as cniht).
At the hour with a silent <h>
Unlike the <k> in words with an initial <kn>, the <h> at the beginning of English words is not predictably silent. While silent in words like ‘hour’, ‘honest’ and ‘heir’, in present-day English, it is generally heard in words like ‘human’, ‘horrible’, ‘history’ and ‘hysterical’, and inconsistently heard in words like ‘herb’. The silent <h> is a relic of French ancestry. Being mostly borrowings from French, where in French, word-initial <h> is unpronounced, it is unsurprising that when <h>-initial words first made their way into English, the <h> was kept silent. However, the visual presence of <h> seems sufficient to prompt its selective sounding over time; see MacMahon (1992, p. 384) and also Oxford Dictionaries’ take on words with the silent <h>.
Typically, we would find the sounding of <h> when we also find <h>-initial words paired with the English article ‘a’ (sounded as a vowel /ə/ or /eɪ/ with no following consonant) instead of ‘an’ (usually sounded in text as a vowel /ə/ followed by a consonant /n/). An example is saying ‘a human’ instead of ‘an human’. This is because where a word-initial <h> remains silent, the first sound of the word in speech is typically a vowel, e.g. /aʊ/ in ‘hour’ /aʊə/ and /ɒ/ in ‘honest’ /ɒnɪst/. And it is these vowels that seemingly justify a pairing of the <h> word with ‘an’, e.g. ‘an hour’.
The consonant /n/ (in ‘an’) averts the dispreferred scenario of a hiatus, which is the continuous sounding of vowels with no consonant to break the continuity, e.g. the continuous sounding of vowels /ə/ and /aʊ/ in ‘a hour’ /əaʊə/. Insofar as this goes, where we find ‘a’ instead of ‘an’ with <h>-initial words, we possibly also have word-initial <h> sounding. In theory, ‘an’ need not be there to avert a hiatus where there is none; and there would be no hiatus if a continuous sounding of vowels involving ‘a’ and the first vowel in a <h> word is broken by the consonant sound of <h>.
Indeed, where the <h> is silent in ‘herb’, ‘herb’ is typically heard with ‘an’ (as is the case in general of English spoken in the United States), and where the <h> in ‘herb’ is sounded, the word is paired with ‘a’ (as is the case with British English in general). There are occasional exceptions to what is said here of course, with ‘an hotel’ and ‘an historical’ being the norm for some speakers even as they sound the <h> in ‘hotel’ and ‘historical’. While these exceptions are worth further study, it remains true to say that a silent (or subsequently sounded) word-initial <h> reveals much not only of the ancestry of the relevant words, but how deeply these words are entwined with the phonological dynamics of English.
The silent <e> that keeps mate and mat sounding different
A commonality between the silent <h> and the silent <k> is that sometimes, they help to differentiate in writing homophones, words that sound exactly the same despite having rather different meanings. ‘Knight’ and ‘night’, for example, sound the same, and we can tell them apart in script only because of the silent <k>. The is the case too for ‘hour’ and ‘our’, where a silent <h> is the only feature that differentiates two identically-sounding words.
There are some silent letters in English words, however, which are not merely there in script, but which also signal pronunciation differences. One of these is the silent <e>. The reason that the silent <e> comes to affect pronunciation is historical. In short, the sounding of the word-final <e> in ‘nose’ and ‘same’ was at some point dropped, consequently allowing the soundless <e> to signal instead that the vowel in the middle of the word it ends is long so that the removal in writing of this <e> indicates a relatively shorter middle vowel (Lass, 1992, p. 82).
An admittedly simplistic, though useful, way of visualising this is how the middle vowel in both ‘nose’ and ‘same’ is represented with two symbols. That is, for ‘nose’, the middle vowel is /əʊ/ (combination of symbols /ə/ and /ʊ/), and for ‘same’, the middle vowel is /eɪ/ (combination of symbols /e/ and /ɪ/).
Compared to the middle vowel in, say, the name ‘Sam’, which is just one symbol /æ/ instead of a combination of symbols, the middle vowels in ‘nose’ and ‘same’ are certainly longer. And it is in this way that a silent word-final <e> indicates a pronunciation difference in, for example, ‘mate’ /meɪt/ (relatively longer middle vowel comprising /e/ and /ɪ/) versus ‘mat’ /mæt/ (relatively shorter middle vowel comprising just /æ/). Other examples include ‘rate’ /reɪt/ versus ‘rat’ /ræt/, and ‘hate’ /heɪt/ versus ‘hat’ /hæt/. To sum up, the word-final <e> in all of ‘same’, ‘mate’, ‘rate’, and ‘hate’, though silent, is a crucial device for setting up pronunciation differences that might not otherwise exist.
Silent word-final <e> (Relatively longer middle vowel) 
Without word-final <e> (Relatively shorter middle vowel)
With a sustained and ever-increasing interest in the study of the sounds of English, script-driven studies of the language may have been somewhat sidelined in recent years. The depth of information revealed in the silent <k>, <h>, and <e> of English words suggests, however, that sometimes, sounds might best be understood in scripts left unspoken. I am strangely reminded of a line of lyrics (from Roman Keating’s song): “You say it best when you say nothing at all”.
 Thanks to Laurie Bauer for alerting me to these examples.
 Thanks to Natalie Chang (Unravel Editor) for the idea of using a table to present this.
Campbell, Lyle. (2004). Historical linguistics: An introduction (2nd edn.). Edinburgh: Edinburgh University Press.
Lass, R. (1992). Phonology and morphology. In Norman Blake (Ed.), The Cambridge history of the English language: Volume II, 1066–1476 (pp. 23–155). Cambridge: Cambridge University Press.
MacMahon, M. K. C. (1992). Phonology. In Suzanne Romaine (Ed.), The Cambridge history of the English language: Volume IV, 1776–1997 (pp. 373–535). Cambridge: Cambridge University Press.