7 letters, 5 syllables, geographic name

It's Monday, so let's start the week off right--with the Sunday puzzle from NPR:

This week's challenge: Think of a well-known seven-letter geographical name in a single word that has just two consonants and yet is pronounced in five syllables.

Wow, that's two weeks in a row with a puzzle that breaks the usual transform thing from Class A to get another thing from Class B format! How should we approach this one? What do we need?

• Lexicon
• If we have a sufficiently large list of words, we can iterate through to find any items that are 7-letters and contain exactly 5 vowels.
• "geographical name": I understand this to mean a place name like "Atlanta" or "Borneo".
• Some of the lexicons we've used in the past might be light on proper names, so we may want to look for something called a gazeteer. These days, a gazeteer is usually just a list of place names, but the name comes from the index printed at the end of an atlas.
• Pronunciation dictionary
• Let's imagine our orthographic search returns 1,000 results of 7-letter, 5-vowel names. With such a large list of results, it would be helpful to run these results through a pronunciation dictionary and return only those with 5 syllables.
• More likely, our search is going to return a much smaller number of results, in which case we can skim through them manually to find the real solution to the puzzle.

Well, that's going to be my approach, and I've got a script started on GitHub that might help you out. Leave a comment if you have a different idea. Good luck and I'll be back after the deadline to share my solution!

3 words that look alike but don't rhyme

Hello Puzzlers! Time for this week's Sunday Puzzle from NPR:

This week's challenge: This week's challenge comes from listener Jim Bricker, of Wayland, Mass. Think of three common six-letter words that have vowels in the second and fifth positions. The last five letters of the words are the same. Only the first letters differ. And none of the words rhyme with either of the others. What words are they?

It's nice to see a puzzle that doesn't follow the familiar format (take a thing from Class A and apply a defined transformation to get a thing from Class B).

How should we approach this puzzle and what resources do we need?

• Lexicon
• I'm using this list of the top 20k most frequent English words:
• https://github.com/first20hours/google-10000-english/blob/master/20k.txt
• Pronunciation dictionary
• We need a list of pronunciations of words in the English language, and I'm using this one:
• https://github.com/cmusphinx/cmudict/blob/master/cmudict.dict
• It has 135,154 pronunciations, so it's safe to assume this will cover any words we've found in the 20k word lexicon.
• A python script

We want our script to do the following:

1. Read in the lexicon.
2. Keep only 6-letter words.
3. Keep only those with (orthographic) vowels (a, e, i, o, u, y) in the 2nd and 5th position.
4. Group remaining words into lists where each list contains words that share the same final 5 letters.
5. Keep only the word groups that contain 3 words or more.
6. Iterate through the groups and retrieve each word's pronunciation(s) from the pron dict.
7. For each pronunciation in the group, compare the (phonetic) vowels. (If two pronunciations have distinct vowel sequences, they do not rhyme.) Print out any groups where at least 3 words have distinct vowels sequences. This should give us the solution.

I've done exactly this, and indeed it is returning the solution. It's also returning a few false solutions, like this one:

[('borrow', 'B AA1 R OW2')]

[('morrow', 'M AA1 R OW0'), ('sorrow', 'S AA1 R OW0')]

[('morrow', 'M AO1 R OW0')]

For my money, these pronunciations more or less do rhyme. That may say something about my own dialect, but I'm confident that another solution is better, and I'll share it after the deadline this Thursday. In the meantime, if you'd like to see my script, you can find it here. Good luck and happy puzzling!

--Levi King

Update