So basically human speech processing is a noisy channel which has limited bandwidth. If the language we encode in has a higher rate of compression, we need to

So basically human speech processing is a noisy channel which has limited bandwidth. If the language we encode in has a higher rate of compression, we need to transmit with fewer errors, but if it has a low rate of compression (or even redundancy) we can support a high rate of transmission and correct for more errors. Which is kind of neat. Thank you Claude Shannon!