Scrambling Letters Within a Word

Don Peterson, someonesdad1 at gmail.com, 6 Oct 2009

A number of years ago someone sent me a paragraph similar to the following:

    I cdnuolt blveiee taht I cluod aulaclty  uesdnatnrd waht I was rdanieg.
    The phaonmneal pweor of the hmuan mnid, aoccdrnig  to a rscheearch at
    Cmabrigde Uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod
    are, the olny iprmoatnt tihng is taht the frist and lsat ltteer be in the
    rghit pclae. The rset can be a taotl mses and you can sitll raed it wouthit
    a porbelm.   Tihs is bcuseae the huamn mnid deos not raed ervey lteter by
    istlef, but the wrod as a wlohe. Amzanig huh? yaeh and I awlyas tghuhot
    slpeling was ipmorantt.

This statement intrigued me so much that I wondered if it worked in general. I wrote a program to scramble the internal letters of each word. I tried it on numerous pieces of text and found that it was generally true for sentences composed of short words. However, longer words were more problematic to comprehension. And while there is quite a bit of redundancy in English and one can often guess the meaning of a word from context without unscrambling the letters, often a scrambled noun or verb with lots of letters would not be decipherable, leading to the sentence not being understood. No doubt the ability to read such scrambled text also depends on your ability to unscramble anagrams.

You can play with the script yourself. Here are the results of running it on various pieces of text:

This first paragraph contains simple words:

    I saw a masegse a few yeras ago taht said taht one cuold mix up the
    letrets in between the fisrt and lsat lertets in the wrdos and the
    maegsse wloud sitll mkae sense.  I wrtoe a poythn siprct to tset tihs
    idea on voarius text.

Here is a paragraph with longer words:

    I tinhk the eessnias of raiedng scuh slrcbmead taxetul
    mateiral denepds hlihgy on the avearge legnth of wdros
    in ecah snnetcee.  Tihs sconed praragaph has illnnatoitney
    longer wdors; the sisocpuin is taht it will be hedrar
    to dhepceir.  What's yuor onpoiin?

Here's a chunk of text on a blog:

    In fcat, deispte what the blgeogrs theeeslmvs bivelee, the frtuue of
    lrtieary ctruule does not lie with bolgs -- or at lseat, it suhldon't.
    The bolg from, that mneicslaly of oeabtnrvosis, onniipos, and likns, is not
    well-suteid to witirng auobt lartiurete, and it is no cdiniconcee taht
    trhee is no lareitry bggoelr wtih the aceuidne and inuefncle of the top
    polaiictl brgegols. For one thnig, leturarite is not nwes the way pcoiilts
    is news -- it doesn't ofefr mtpilule evetns erevy day for the bgegolr to
    cmonemt on. For athenor, beziitsed cnaromtemy, whcih is all the bolg form
    awolls, is next to uelsess when it cemos to tkliang aubot bkoos. Ltrrieay
    ccsiritim is olny worth hanvig if it at laset svitres to be liatrrey in its
    own rihgt, wtih a spoce, citopelxmy, and aroithtuy taht no bggleor I know
    even wtnas to acivhee. The only ufusel prat of msot book bgols, in fcat,
    are the links to lnog-from esyass and aictrels by psaeroifosnl werrtis,
    ulsualy from pnirt janrluos.

Here's an abstract of a technical paper:

    5-Oxo-6,8,11,14-eeteitaiasncoorc aicd (5-oxo-ETE) is a rtlcneey
    dveicorsed mtaebliote of aiidhcoranc acid that aacevttis haumn
    npleoiuthrs by a mhecasnim iennnepdedt of the rotcpeer for lienruteoke
    B4 (LTB4). The obvcjeetis of tihs study wree to idtnfeiy the moajr
    mibeaotelts of 5-oxo-ETE in nhlroutpies and to cmraope the biilgooc
    aviieitcts of 5-oxo-ETE wtih thsoe of its matieeotbls and oehtr 5-
    onixascdiooeos.  Npuoirlehts rldaipy cneeotvrd 5-oxo-ETE to its ogema-
    ooaitdixn poudcrt, 5-oxo-20-hrdoxyy-6E,8Z,11Z,14Z- eiorctsenaiateoc
    aicd. This copuonmd was narley 100 tmies lses pontet tahn 5-oxo-ETE in
    eevaitnlg cioylotsc cacuilm levels in nehultopirs. Mhytoletain of the
    cayxbrol group of 5-oxo-ETE rlteused in a 20-flod loss of ptnceoy,
    whreaes rpeeecmnalt of the 8,9-cis double bnod by a tarns dbloue bond
    recuded ptecnoy by aoubt sioxlfd.  Siilmar rsuelts were obtaenid for
    the eftcfes of the abvoe coumpodns on nreitophul mtairoign. 5-Oxo-20-
    hrxyody-6E,8Z,11Z,14Z- eonieicartaesotc aicd, 5-oxo-8-tanrs-ETE, and 5-
    oxo-ETE meythl eestr deetizsisned nhiluproets to 5-oxo-ETE. 5-Oxo-ETE-
    iunedcd cclauim miaozlibiton was ibhnteiid by peenteamrrtt of the cells
    wtih ptrsiesus txion. 5-Oxo mbieetoalts of 6-trnas-LTB4 and 12-epi-6-
    tarns-LTB4 had waek siotralmuty etffces on cuiaclm lvlees and moagirtin
    taht aapeperd to be meetadid pilaimrry by simaolitutn of LTB4
    rocretpes. Thsee stuieds iidtcnae taht the 5-oxo gourp, the oemga-end
    of the mlclueoe, and the cboyxral gorup are all itnrampot for the
    bologiic aivictty of 5-oxo-ETE, which may be meeditad by a G proetin-
    lenkid rctopeer. The bgiolioc avicttiy of 5-oxo-ETE can be teamtirned
    by omega-oixoatdin.

Here's the python script:

'''
Script to scramble the letters in words, leaving the first and last letter
the same.

Each file on the command line is processed in sequence, sending the
processed text to stdout.  If no files are given, stdin is read.
'''

import sys, random
out = sys.stdout.write

def ProcessWord(word):
    if len(word) < 4:
        return word
    first_letter, last_letter = word[0], word[-1]
    center = list(word[1:-1])
    random.shuffle(center)
    return first_letter + ''.join(center) + last_letter

def ProcessString(s):
    word, in_word = "", False
    for char in s:
        if char.isalpha():
            in_word = True
            word += char
            continue
        else:
            if in_word:
                word = ProcessWord(word)
                out(word)
                word, in_word = "", False
            out(char)

def main():
    if len(sys.argv) == 1:
        s = sys.stdin.read()
        ProcessString(s)
    else:
        for file in sys.argv[1:]:
            s = open(file).read()
            ProcessString(s)
    exit(0)

main()