Language is complex, to say the least. The huge variation in human speech patterns makes creating effective IVR systems difficult, but at the end of the day it is this complexity that makes us unique.
Although scientists continue to discover seeming “animal languages” (bird song, whale calls, dolphins’ clicks, etc.), humans by far have the most complex language of any creature on this planet.
IVR systems break down language and convert it into a vast array of data so the system can analyze and interpret what we’re trying to say. The complexity of language, however, makes this difficult.
By the University of Oregon’s last count, there are about 40 different phonemes (or sound units) in the English language alone, although some languages have far more.
Recognizing words and sentences from the huge number of possible phoneme combinations is one thing, but as we all know, English likes to make it even more complicated.
For one, there are about a million different ways to say “yes” in English: yes, yup, yeah, yep, mmhmm, yes’m, yessir, uh-huh…the list goes on forever. Add in all of the improper contractions like “gonna” and “dunno” that pepper our conversation and you suddenly have quite a large list of words.
IVR systems have to account for all of those synonyms and then further account for regional dialects. The system has to know that when a Bostonite and a Texan each say, “car,” that they both mean the same thing.
Finally, the system has to recognize words no matter the pitch, tone of voice or emotion.
Suddenly IVR systems seem much more impressive.
What’s even more impressive, though, is how science has managed to give the ability to speak back to those who have lost it. They do this through a huge number of devices and programs, and now, through surgery.
…Stay tuned for the rest in “Pioneering Speech Surgery”…