Search Ace Linguist

July 21, 2020

I Dream(p)t of Euphonic Insertion - or Phonetic Intrusion

You may be familiar with the question of which is the "proper" past tense form of the verb 'dream': is it 'dreamed' or 'dreamt'? But one form has long since dropped out of discussion: 'dreampt,' a form old enough to be found in Shakespeare's plays. The form 'dreampt' seems to survive only for people invoking an old-fashioned mystique. But whence dreampt - where did that 'p' in 'dreampt' come from?

Romeo: I dreampt a dreame to night. Mer: And so did I. Rom: Well what was yours? Mer: That dreamers often lye.

The 'p' in 'dreampt' was introduced through a phonetic process called 'intrusion due to coarticulation,' or 'euphonic insertion' in the older philological tradition. Intrusion happens due to a fact about how vocal sounds are made. You see, the tongue and lips have to actually move from one place to another when we are speaking. When we write 'dreamt', or [drɛmt] in IPA, the letters abstract away this fact. The cluster [mt] makes us think that the velum lowers to produce a nasal sound, the lips close perfectly, the velum raises and the lips open, and then the tongue strikes the alveolar ridge.

But this sequence is an ideal version of [mt]. The reality is more often as follows, using the example of 'something'. Bold emphasis and paragraph breaks are included for emphasis and readability. (Reetz, 2009)

[...] Intrusions [are] when phones are inserted. These insertions can [...] result from coarticulation. In English, this happens when a nasal consonant precedes a voiceless fricative, as in the word 'something.' In these cases, a voiceless plosive with the same place as the nasal may intrude between the nasal and the voiceless fricative.

Namely, to produce the nasal, the oral pathway has to be closed completely (similar to an oral stop) and the velum is lowered.

Next, in the production of the following fricative, the velum is raised and the plosive closure is released at the same time to allow turbulent airflow for the fricative.

If the velum is closed first and the oral closure is released slightly later, the articulation is the same as an oral stop: an oral closure with raised velum and a release. As a result, an oral stop has been inserted.

For example, chance [t͡ʃæ̃ns] can become [t͡ʃæ̃nts], length [lɛ̃ŋθ] can become [lɛ̃ŋkθ], or something [sʌ̃mθɪ̃ŋ] can become [sʌ̃mpθɪ̃ŋ].

The quote above focuses on nasals followed by fricatives, but the same can happen to nasals followed by stops in a different articulation.

Indeed, English has a rule that nasals + stops must shared the same place of articulation (in other words, they must be homorganic consonants) within the same morpheme. Think of how strange the sequence [anp] would be in English. I bet when pronouncing it, you want to turn that [n] into an [m] to create the comfortable [amp].

'Dreamt' could be analyzed as two morphemes: 'drem' plus the past tense marker '-t', which explains why this non-homorganic (or heterorganic) sequence exists in English.

Em(p)ty Movement

'Dreampt' is not the only word to have a 'p' sound euphonically inserted. In some words, the intrusive 'p' became part of the standard spelling, and therefore part of the standard pronunciation. One such word is 'empty,' which originated as Old English 'æmettig.' Once that middle vowel was lost, the spelling 'empty' with a 'p' appears to have proliferated.

Some spellings came to include this 'p', and other spellings do not include them. H.L. Mencken noted:

Boughten and dreampt present greater difficulties. [...] The p-sound in dreampt follows a phonetic law that is also seen in warm(p)th, com(p)fort, and some(p)thing, and that has actually inserted a p in Thompson (=Tom’s son).

This sort of intrusion occurs cross-linguistically as well. The word 'tempt' in English comes from Latin (via French) 'temptare,' a variation of 'tentare.' It's not clear how that 'n' became an 'm' in the first place, but we can imagine that once someone attempted to pronounce 'temtare,' that dastardly 'p' moved in to create 'temptare.'

There are other types of intrusion other than a [p] being inserted between an [m] and a fricative or alveolar stop. As mentioned above, 'chance' can become 'chants,' and 'length' can become 'lenkth'. The Middle English compendium does have a single citation for 'lenkthe'. I couldn't find a word with [ns] that was spelled 'nts', but I also did not look very far.

Though the spelling 'printce' for 'prince' doesn't seem to exist in Middle Enlish or in today's modern English, the similarity between these two words has not gone unnoticed. The following joke from the Animaniacs series depends entirely on the similarity between 'prints' and 'Prince', as well as the flexibility of 'finger' as a noun and 'finger' as a verb.

Dot: I found Prince! [holding up Prince, the singer]
Yakko: No no no, finger prints!
Dot: I don't think so.

Works Cited


  1. This is fascinating. Why was the p dropped from the spelling of dreamt though? If some words were changed due to the pronunciation allowing a 'p' sound to slip in, then why did "dreampt" have the reverse effect?

  2. Thank you for this. Excellent analysis, I enjoyed this entirely too much.