Negative middle ear pressure presents something of a paradox. The ‘Type C’ tympanogram, in which the peak of the tympanogram occurs below zero pressure, seems to indicate that the air pressure in the middle ear is actually below atmospheric pressure – that there is a degree of suction – and yet the peak can remain persistently in place even if the subject swallows and opens their Eustachian tube. Negative middle ear pressure can even be measured when a subject has a permanently open (patulous) Eustachian tube, a situation that seems physically impossible. This paper reviews the paradox and concludes that in many cases of “negative middle ear pressure” the actual pressure inside the middle ear is in fact zero, but the tympanometric offset comes about because of the unappreciated action of the tensor tympani: when this muscle contracts, it pulls the eardrum inwards, and this inwards force is registered as negative middle ear pressure during tympanometry. That is, the force exerted by the muscle needs to be countered by a negative pressure in the ear canal in order to bring the eardrum back to its equilibrium position. This interpretation is reinforced by a number of findings in the literature, which are reviewed. A proposal for how tensor tympani effects might be separated from actual middle ear pressure offsets is made.