Preservation of primary phonetic and acoustic cues of phonemes trigger their perceptual identification. Time reversal of speech both preserves and alters phonetic and acoustic features of speech signals. Invariant features such the power spectra of a signal are usually maintained whilst properties such as duration and the shape of the temporal envelope, as well as finer details of the acoustic spectrum are altered (Grataloup, Hoen, Veuillet, Collet, Pellegrino and Meunier, 2009). Non-continuant speech sounds are more susceptible to altered perception in reversals as assymetry typically occurs in the shape of the temporal envelope. This is the case in stop bursts, abrupt vowel onsets, and ramping (smooth increase in amplitude) and damping (smooth decay) of signals (Pellegrino, Ferragne and Meunier, 2010). Time reversal of these features alter the characteristics of the speech signal, permitting perception of alternative phonemes, and even the addition of phonemes to the speech signal, or the omission of phonemes from the forward speech.

One study has been conducted that investigates the preservation of phonetic cues in time reversed speech and the perception of reversed phonemes. Pellegrino, Ferragne and Meunier (2010) conducted an experiment which required four phoneticians to listen to pseudowords that were recorded and played in reverse, and phonemically transcribe what they heard. The results of the study showed that around 25% of the original segments from the forward speech were exactly retrieved in reverse. The experiment also demonstrated that certain phoneme types were more likely to be distinguished than others. Fricatives (e.g. /f, v/) liquids (e.g. /l/) and nasals (e.g. /n, m/) were identified at a rate above 90%, and vowels at close to 90%. The authors suggest that the high rate of identification likely reflects the invariance of continuant waveforms preserving a high level of perceptual cues permitting perception. Rhotics (e.g. /r/) and voiced stops (e.g. /b, d, g/) were identified at an intermediate level (66.7% and 61.8% respectively). Listeners, however, were inaccurate with unvoiced stops (e.g. /p, k, t/), with a rate of only 9.4%, as well as schwas (mid central neutral vowel /ə/).

The ones that were not correctly recognised were identified as phonemes having alternative place and/or manner of articulation. 30% of unvoiced stops were transcribed as fricatives. 25% were identified as stops, which also included other stop types such as glottal stops or unreleased voiced stops. 28% were heard as a cluster; for example, a final /t/ in the natural speech was heard as an /sn/ cluster. The authors suggest that the /n/ arose from the ramping of the vowel in the time-reversal signal. 7% were transcribed as a sonorant (r, l, m, n, w, y) while 10% of the stop segments were not detected.

The findings of this study suggest that not only are speech sounds from the forward speech heard in reverse, sounds that are not in the forward speech are also perceived as phonemes.

These perceptions are typical in Reverse Speech. Although many phonemes from the forward speech are perceived, others are heard as alternative sounds, and this is certainly the case with unvoiced stops. They can be perceived as a phoneme with a different place of articulation (e.g. /t/ → /k/ or different manner of articulation (e.g. /t/ →/s/, /p/ → /f/). An alveolar stop and alveolar /l/ can convert into another alveolar consonant; for example, /t/ or /d/ may be perceived as /n/ or vice versa. Others may be heard as allophones (different variation of the one phoneme; e.g. /t/ → /ʔ/ or unreleased /t/), or a similar phoneme such as an alveolar tap /ɾ/. Phoneme addition can occur such as /t/→ /st/. Stop bursts can disappear when reversed, lost in the vowel sound that came before it in reverse, resulting in perception of an alternative phoneme, an unreleased allophone, or omission altogether. Omission of sounds from the forward speech is a common occurrence. Light articulation of consonants or the strong frication of vowels next to a consonant may result in non-recognition of the consonant.

Some sounds in time reversed speech are highly ambiguous and may be heard differently by different listeners. Alteration of phonemic cues through reversing or degrading of the sound through audio noise or poor audio quality contribute to ambiguity. In this case, one’s grammatical and lexical knowledge comes into play in phoneme selection, projecting the desired phoneme to produce meaning.

Reverse Speech is very much about the perception of speech sounds and finding meaning though the building of strings of language that make some grammatical and syntactical sense. But of course, this is very much the case for normal speech as well. We turn the sounds uttered by another into coherent meaning. When listening to speech, we cannot actually perceive each individual speech sound. We assume that they are there. However, if we were to examine the individual segments of spontaneous forward speech, we would find that not all phonemes of the heard words are recognisable; they may sound different or be missing altogether. Yet, there is ample remaining of the speech signal to perceive a coherent string of words. The rest is projected into it.

So, we can now see that Reverse Speech is composed of perceivable phonemes and segments. Not covered by Pellegrino et al. is whether the segments produce lexical information. It can be easily proven that they indeed do. However, to perceive strings of language correctly, one needs to operate within linguistic possibilities and parameters. This entails examination of phonemes and segments of reversed speech as well as comparing them to the information in the forward speech. This means understanding linguistic processes. This also means knowing that some speech sounds in forward speech can be heard differently to the sounds which normally make up words.  It is important to know what is wrong with the string of words just as it is important to know what is right. This helps to set reasonable linguistic parameters for what can be accepted as linguistically viable. There are innumerable examples out there in ‘Reverse Speech World’ that are obviously not what they are claimed to say. There are also many that can sound like what they are attested to be, yet still lack the necessary evidence for it.

Yet, strings do occur that mirror acceptable language. Nevertheless, proving that they are anything but coincidental is another matter. Every day, there are perhaps trillions of strings of language produced by speakers around the world. Quite naturally, ‘words’ will appear that are purely coincidental, even if they are a grammatically acceptable string of two, three or four words which are composed of perhaps one or two content words and one or two particles. One can shake these in front of linguistics all day and get a response like “that’s interesting, but no cigar!”, even if they did seem to have some meaning regarding the speaker and what he was saying. For attention to be garnered, linguistically viable strings that are much longer need to occur; say, a minimum of 7 words in length with ample examples of ones that are more than 10 words and even as long as 15 -20 words.

Funnily enough, they exist.


Grataloup, C., Hoen, M., Veuillet, E., Collet, L., Pellegrino, F & Meunier, F. (2009). Speech Restoration: An Interactive Process, Journal of Speech, Language, and Hearing Research, 52, 827-838.
Pellegrino, F., Ferragne, E., & Meunier, F. (2010). 2010, a speech oddity: Phonetic transcription of reversed speech. Interspeech 2010, 1221 – 1224.

I have taken clusters of language-like words from the reversed speech of a participant at a Reverse Speech seminar and presented them here with an examination of the linguistics and discussion of possible meaning. That I have documented reversals, does not mean I believe that they are real. Some of them can be heard in different ways, and some are not sound enough phonemically. In fact some are quite inane. Audio noise also plays a role in influencing perception here. It is a look at what someone might hear, document and present as Reverse Speech, and an exploration of what may ‘pass the grade’ as Reverse Speech. 

RS: in/and that fuck the raper  

FS: I get moved how um people can be so down and they (?) it just becomes ah joyful and happy and they see a different, their skin changes and they become [happy and confident ‘n i]t it just moves me to see when ….

The raper is behind [happy an]. There is a sense of an /r/ in the RS before articulation of /n/ occurs in the FS, but this could be aided by the audio noise. /n/ produces an approximation of the. Fuck comes from [conf]ident where /n/ is greatly decayed. A northern English [u] vowel occurs (note: speaker is originally from Yorkshire). In/and that comes from confi[dent ‘n i]t.


This is fairly sound linguistically, assuming audio noise did not play too great a part in it. Whether there was a real raper is something that would need to be determined (and also whether it has something to do with the potential words nigga/Muhammed, which occur close by, analysed further below). In the FS there is a great wish for positive, happy experiences, but in the reverse, there is something quite different going on.


RS: Watch me kill one day  

FS: [Yeah, I work in Tro]pic and I’m the senior manager in Tropics.

[o]ne comes from [I w]ork,  /n/ comes at the onset of [I] in the FS (it is possible that the speaker uttered [‘n] before [I] in the FS but the noise masks it). In the FS at /d/ in [day] David Oates speaks at the same time. It sounds like a soft /t/ comes from the speaker and the release of the vowel from David Oates can give the impression of a /d/Kill comes at [work]. There is a sense of [m]e at i[n], but this may be aided by audio noise. Watch comes from [tro]. The rounded vowel combined with the mouth moving towards /p/ produces /w/. tro[pic] produces kep in the RS. There is an assumption the reversal begins after this.


Of course audio noise hinders identification. [day] is uncertain, and there is an assumption of [me] 2) a syllable – kep (tro[pic]) – has not been included. It is not known whether this is something that should be a part of it.

Interestingly, David Oates says in reverse after this “Oh, right”.


RS: I had this attack 

FS: It’s not about money it’s not about possessions it’s not about anything li[ke that, it’s abou]t changing people’s being

The speaker’s emotions produced the /h/, allowing for perception of [I ha]. One can quite easily perceive /d/ at the point of /b/ in FS [about], however, /b/ can still be perceived. One can get a sense of [th] from the constriction at the release of /a/ and onset of /b/ in [ab]out combined with the frication occurring.  Attack comes from li[ke that i]t’s. The FS [th] is decayed and /k/ is heard.


The final in [had] lets this reversal down a little as it is dubious. Perhaps it is meant to be have, perhaps it is gibberish. A very light [th] is not an issue as this can occur in FS.  Otherwise, the stress pattern is appropriate. Based on this, one will need to decide if there is enough that is right about it. Assuming accuracy, one would then need to consider whether it is a negative emotional/psychological event that occurs or occurred to her, or whether it is a physical one, and this may tie it into other reversals documented here.


RS: I kiss Noah. See a loss, some deliverer I want  

FS: Um what’s stopping me from be[ing who I really, my soul is I want to ge]t everything you know I just …

I kiss – t[o get]  /o/ gives sense of I; /k/ from /g/ and /s/ from sibilance of /t/ is reverse.

Noah – [I want] /n/ from /n/; [oa] comes from [I wa]

See a loss – [soul is] [see] from [is]; [a] from release of /l/ moving towards [i]; [loss] from [soul].

some – [my s]oul

deliverer – [I really] /d/ is perceived the release of [y]; sense of /v/ as well as /r/ comes from [I] and /r/ together; sense of an extended [er] comes from including enough of [I]. As a FS version of this word would often not be enunciated clearly in regard to [erer] unless spoken slowly, I find the possibility of [deliverer] reasonable here.

I want – [ing who I] The remaining part of I (not utilised in the [er] above) goes toward producing RS I. /w/ comes from the rounded vowel in [who], and /n/ comes from [ng].


Kiss Noah and deliverer may indicate a strong desire for someone to help her and guide her. The reversal could then be a response to the consultation with David Oates.


This is a long language-like string. Here is an approximation of words.

RS: You/he/it had/got me/my head swimming – it’s true – then head him off – Australian on eggs will beat my egg through 

FS: I work in Tropic, and I’m a senior manager in Tropics (?) and sell (vegan) skin care and make up which I just [love getting on people’s skin and they learnt it from me then, then it just and it moves me when I] can see them change and transform.

(a) RS: You/he/it had/got me/my head swimming 

FS: and it moves me when I

you/he/it had/got – whe[n I]  The emotionality and duration of I creates an aspirated /h/ or, alternatively, the stop nature of the emotional noise can create sense of a [g] /d/ comes at FS /n/The pronoun is uncertain.  

me – m[e wh] The /w/ produces perception of /m/

head – [me] The speaker’s emotionality creates the /h/. Sense of unreleased /d/ occurs at the onset of FS /e/.

swimming – [an’ it moves]  swim comes from moves with /v/ allowing perception of /w/ (but also /l/).  ing comes from a[n’i]t.

(b) RS: it’s true

FS: it just

ju[st] allows a sense of i[t’s] and a further sound produced by the speaker after [just]  allows a sense of [i]t’sThe [j] permits a sense of [tr]ue, however, there is no /r/,  and therefore this is uncertain.

(c) RS: then head him off 

FS: from me then then

then head [then then] This is just one alternative; it is too unclear to be certain of any words.

him off  [from me]

(d) RS: Australian on eggs will beat my egg through 

FS: [getting on people’s skin and they learnt i]t

Aus – learn[t i]t /t/ produces /s/

tra – l[earn t]o – there is an /n/ in the RS rather than /t/.

lian  – [nd they l]earn [th] disappears; /a/ is pronounced rather than a schwa (unpronounced vowel).


on eggs – [skin a]nd  The vowel is indicative of eggs, however the softness of the consonant could make it ex or even aches. 

will beat – [people]’s  a sense of an unreleased /t/ can be perceived at initial /p/ in [people]. /w/ can be perceived at [le].

my egg – [getting on] /m/ is perceived at /n/ as it assimilates to the [p] in the FS; [y] comes from the nasalised /o/ + /i/ in the FS) [ng] disappears and /t/ is glottal like and cannot be heard in the RS); egg is from [ge]tting

through – [love] – frication from /v/ gives sense of [thr]


Of course, this is just one way to hear it.  If there is actual language occurring, some of it is quite likely something else. Her head swimming may be a reaction to the situation she found herself in, speaking to David Oates about herself in front of people; the fact that Oates is Australian makes the closeness to ‘Australian’ interesting. Some of it is certainly rather inane. Just because it sounds like language doesn’t make it language.


RS:  Saw what happened, idea for China; yeah this decision the end of my bills/Soul wake up, an idea for China, yeah this decision the end of my bills 

FS: I want to be able to just not [to worry about money and just to be financially free again like I was].

saw/soul – [was] /w/ can create a sense of /l/
what/wake– li[ke I w]as Although /t/ is not directly evident this could just be elided. Alternatively, there remains some sense of /k/ for a possible [wake].
happened/up – agai[nlik]e The aspiration and weakness of RS /k/ can give a sense of /h/, /p/ comes at FS [l]. If the reversal is shortened here, and the /n/ used as yej start of the next reversal segment, it can be [up].

Yeah this decision –  a[nd just to be] [just] gives [cis] in [decision]. Although /d/ is not fully formed, there is a sense of an attempted inclusion of a stop at the point of FS [to]. In FS, this is also the case. 

the end of my – [about money an]d  [th] naturally assimilates to the /n/ in [decision]. /b/ disappears in the RS as does /t/ which had assimilated to /m/ anyway in the FS, therefore m[y] occurs at [about] and  /m/ comes from /m/.

bills – [to worry] an indeterminant sound occurs at the FS release of /y/ allowing some impression of a possible /b/; /ll/ comes from [worr] and /s/ comes from [t]o.


The ending of [decision] can be lost in an overall listening; however, it is quite possible. The reversal seems to say that an idea in regard to China can be fruitful for her financially.


RS: We’re ready it would/Edward (b)ook you for (the) Alice, Alison, (b)ook (m)e for (the) Alice 

FS: an’ to let, to let (other?) people would be a real gift

We’re – [real]  soft /l/ produces /w/ in RS

ready – b[e a r]eal /d/ comes from transition of FS /e/ to /a/ where the tongue is high and forward.

it would/Edward – [would be]  i[t] or E[d] can be perceived at [d b]; /w/ comes from the rounded vowel w[oul]d; /d/ can be perceived at [w]ould.

(b)ook – peo[ple]  A stop /k/ can be perceived at /p/. This is unclear as there sounds like a double consonant at the beginning where one is /l/ and the other /b/ or /d/.

you – oth[er peo]ple  high front vowel allows perception of /y/, and movement of vowel toward /p/ in the FS allows for y[ou].

for (the) [oth]er –  The frication at [th] gives a sense of /f/. The change in intensity between [o] and [th] in the FS gives a sense of [the] following [for], which may not be heard in a more general listening. Here is the The Alice with for not included

Alice – [to le]t There is a pause between Alice and Alison where an unreleased /t/ occurs. Although nothing can be heard, one could also make an assumption that it is Alison.

Alison – [an’ to le]t

(b)ook – peo[ple]  A [k] final is perceived. One can hear [l] as the initial; [b] can easily be heard due in part to the sudden ramping up of energy.

(m)e [peo]ple Initial is actually [p] which comes from the ramping up energy of the vowel [eo]. The FS [p] disappears.

for (the)  [other]  A consonant such as [p] or [f] may be heard o the release of [er] in the FS. There is sense of [the] from [oth] which may not be heard in a general listening.

Alice  [to let] Alice comes from [to le] with the [t] producing [s]. The other [t] in the FS disappears.


Of course this is one way to hear it. Alice is repeated from the FS words [to let], which could just be seen as coincidence; however, one may argue that the subconscious has influenced use of the words.  (b)ook is an issue due to the [l] that can also be heard. It is promotes some interest, however, due to [we’re ready] and [(the) Alice] and the possible book, in view of the speaker talking with David Oates, an Australian, and The Alice being a stereotypical place to visit for overseas tourists.


RS: Put the self-pride, Earls and it numb it, (put) souls in there, (nurse/there?) she/you can’t solve wrecker 

FS: I was very abundant and then um [I kind of lost my passion and lost and went into a dark place (laughter)]. And um I want people to be walking around happy and joyful.

put the – [laughter] this comes at the laughter, could be heard as with the, also. If this is ignored, the reversal begins at self-pride

self pride – [dark place] /p/ can be perceived at FS /k/. /r/ is not really there, so it is assumed. The /f/ is not particularly evident, and it can sound like sell; however, we can assume assimilation of /f/ to [p] 

Earls – [to a] /t/ provides the sibilant as is typical in RS.

an’ it – wen[t in]

numb it – [an’ wen)t a[n] gives a sense of /m/. 

put souls – [lost] /t/ produces sense of the.

in there – [and]  

nurse she/you can’t/ – los[t my passion] – an approximation of nurse occurs. Continuation of /sh/ + vowel allows for some sense of she, but perceived rounded can give impression of you with the sibilant fully part of nurse. With can’t, [c] may be perceived from release of m[y], although this could also be /p/; /m/ may be heard as /n/, and natural ending of can’t can be heard with FS /t/.

solve – o[f lost]

wrecker – [I kind o]f  [nd] decays to /r/; rounded [o] produces /w/. 


Most of the words are reasonable, self-pride is a bit iffy. Perhaps the two put may be meant to be something else. Nurse is doubtful, and this may need to be dropped.  If there is anything in this, we see aspects of her psychology – self-pride, souls, Earls, numb it (self pride or souls?), can’t solve the wrecker. Earls only makes sense if it is accepted as a metaphor for a leading aspect of the Self.


RS: an’ if it’s (a) certain nigga, Muhammed 

FS: I’d be singing down the street singing (?) [with my handbag and just skippin’ a]long

an’ if it’s comes behind [skippin’a]long. The stop effect of the FS /p/ disappears, allowing perception of /f/. /k/ can be heard in i[t]’s, however, similar acoustic cues allow for some perception of /t/ and the influence of FS /s/ in creating a RS /t/ dominates the /k/. The sibilant /s/ is longer than normal. It is possible to perceive [a] following [it’s]. The reason for this is between the /s/ of it’[s] and [c]ertain at jus[t] skipping -though it continues to be a sibilant, intensity drops, emulating someone who doesn’t bother releasing /s/ in order to articulate /a/. certain comes from a[nd just]. The stop component of the affricate [j] allows for /t/ perception. The lengthened [n] from [and] in the FS gives both the [n] final in [certain] and [n] initial in [nigga]Nigga comes from handb[ag an]d, where the extended /n/ in the FS allows for a separate perception of /n/ from certai[n]. The vowel tends to fall into a diphthong /ei/ and the final syllable [ga] tends to have higher stress, so the word may sound a little odd. The stress pattern is interesting, however. The extended /s/ in it’s can occur when someone is hesitating for a moment. Certain drops in intensity as if the speaker knows they could be saying something wrong. The higher stress on ni[gga] could be an intonation pattern to indicate the speaker is saying something important and you should listen. Muhammed comes at [with my handb]ag. The [nd] assimilates to the [b], and in reverse sounds like /m/. Muhamm[ed] occurs at [with]. Closer listening will give a perception of did at [ed]. Therefore, the word is iffy. 


In it’s, /t/ could be /k/. Nigga, although phonemes are there, is articulated a little strangely, and there is a [d] or [th] after the mid /m/ in Muhammed, which can get ignored quite easily. These are the issues with it. That known, one would then need to decide whether there is a reversal occurring or just coincidental sounds.


Immediately following the above reversal in the reverse direction is this:

RS: Slip I hang a (spirit/skirt?), and/in lightning 

FS: I’d be just si[nging down the street singing (happily?)] with my handbag and just skippin’ along ….

Slip is behind what I have assumed might be [happily] in the FS, though it sounds like [hapils].  I hang a is behind s[inging ha]. RS ng is at FS ng. I comes from h[a]. What I noted as skirt/spirit is behind [street s]. The [t s] produces the initial /s/. There is sense of an ambiguous stop consonant from the release of the vowel str[ee]t; one could hear it as /k/, /p/, /t/, or even a /y/ with a hard onset. The final /t/ is produced by [st]reet. Sense of and/in comes from dow[n the]. Sense of lightning comes from si[nging down]. However, it is actually like the sounds for nightding; it is just easy to project the required sounds into it.

There are sounds that follow this, but occur at a different tempo. If they were included it would sound something like an’ nigh ding a slipper with sl a bit distorted.


This is messy and quite inane. Although we can project lightning into it, evidence doesn’t support it.


Byrne and Normand. (2000). The Demon-Haunted Sentence: A Skeptical Analysis of Reverse Speech

The authors make claims that can lead readers to believe that David Oates is a shyster, fraudently profiteering from his claims. For example:

  • ‘His company … is dedicated to profiting from his discovery’.
  • A ‘reviewer’ of David Oates’ book, Susan Brombacher, concluded ‘that (Oates) seems more interested in making a profit than educating others.
  • ‘The Reverse Speech Web page contains a plethora of merchandise and services available to consumers at considerable prices’.
  • ‘ ….. those who pay (David Oates) a hefty sum and go through his training can then serve as expert witnesses and command hefty sums themselves’.

It is my suggestion that researchers like Byrne and Normand drop the fact that David Oates makes an income from Reverse Speech, and avoid the urge to use this as a method to discredit the man and the phenomenon.

The attention given to the fact that Reverse Speech is an income-producing enterprise takes away from the necessary argument concerning its genuineness, appears as small-mindedness, and smacks of academic ‘poor form’.  They would do well in ascertaining whether real speech exists, and what it is that is producing it.

The authors write:

“The person trained to hear reverse messages could intentionally or unintentionally report that speech contains hidden incriminating evidence. Many people are not prepared to refute such contrived evidence”.

The authors state that the potential for harm is enormous if Reverse Speech is accepted in places such as a court of law. I agree that it is problematic. It is easy to err without sound skills. This goes for any use of Reverse Speech. A certain perceived stress pattern and the existence of some phonemes can cause the listener to hear a phrase even though other phonemes are incorrect, ambiguous or missing. The ambiguities can mean that it lacks evidence as the documented reversal, and could easily be heard as something else, or nothing at all. Of course, in normal speech, all kinds of processes occur – phoneme assimilations, elision, dropped phonemes and so on – one needs to work with a knowledge of language and its processes and what might be viable within a reasonable set of limitations. After an informed analysis leads the analyser to believe that the reversal indeed represents the sounds of language, it needs to be ascertained whether it is genuine unconscious language rather than just coincidental language-like sounds. Next, what/who the language refers to would have to be identified through a cogent interpretation. What aspect of the person is the ‘voice’? Is it an aspect that represents normal conscious thinking or how he/she views events and those around them? Or is it a denied aspect of the self? Or is it an aspect that communicates at a deeper, more profound level? When the voice uses personal pronouns, such as ‘I’, ‘he’ and ‘you’, who is this entity referring to? Is it the speaker himself or a third person, and if so, who is that third person?  Does the reversal refer directly to the spoken forward speech, or is it the indirect result of a trigger from the forward speech, and refer instead to events or memories of the further past?

The authors apply to Reverse Speech circumstances and conditions which can result in misunderstood and misinterpreted language. For example, they refer to B.F. Skinner’s verbal summator. The authors state:

“The verbal summator consisted of a phonograph (or tape) of random vowel sounds that were grouped together in such a way as to not produce any systematic phonetic groupings. These random phonetic sounds were arranged into patterns that approximated common stress patterns in everyday conversation’.

After such strings of nonsense syllables were arranged, they were played for subjects at barely audible volume levels. After repeatedly listening to these sounds, subjects reported “hearing” the phonograph or the tape “say” things. These sentences, or sentence fragments, did not actually exist and, as such, were considered to be utterances that were already strong in the subject’s repertoire. Put another way, they were “projecting” their own thoughts onto the sounds they were hearing”.

The authors go on to add, “phonemes may sound similar to a meaningful phrase but are really sound salad’. A listener expecting to hear a certain phrase will likely do so”.

Quite true, we project into the signal to hear language, and even more so when there are issues of low audibility, background noise, or audio noise occurring. Masking effects cause the listener to hear a particular utterance that is not there. However, it is not only in Reverse Speech that this can occur; it also occurs in normal speech. Projection into sounds and priming or prompting are issues that affect the hearing of speech whether it is forward or backward. However, Reverse Speech is at a distinct disadvantage compared to normal speech. Here is why:

  • When listening to normal speech, we know it is real language (unless someone has played a trick and recorded noises from a dog that mirror a human stress pattern, and used masking noise over it!). With Reverse Speech, speech needs to be identified from non-speech. This means recognising what sounds reasonably constitute language. It also means being able to recognise a structure that sounds like language as being genuine or simply coincidental.
  • The analyst needs to identify the beginning and the end of the reversal. Identifying only some of it can alter how it is interpreted overall, or it can result in completely different words due to how it has been ‘cut’ from a longer utterance. The same thing would occur in normal speech.
  • People are not used to hearing Reverse Speech and, much like hearing an accent or language style one is not used to, it can lead to difficulties in comprehension.
  • We do not understand the mechanisms by which it occurs.

These issues do not mean that Reverse Speech is not a real phenomenon. But it does mean that there are considerable challenges associated with it – many of the difficulties that can occur with normal speech in a comparable situation. Because of the inherent challenges of Reverse Speech, those who investigate it ought to possess appropriate linguistic skills.

In reference to papers such as Byrne and Normand, I suggest that anyone who seeks to offer a critique of the existence of Reverse Speech, do so by first identifying whether there are grammatically-acceptable, linguistically-viable utterances in reverse that are significant in number. I have yet to see a critical study that addresses this basic first point of consideration. From the recognition of the large number of utterances in reverse which are of interest linguistically, one then needs to show whether they are simply coincidental sounds and have nothing to do with  some level of consciousness. An honest appraisal of the not-insignificant amount of material available that is arguably linguistically-viable, as well as grammatically and syntactically acceptable, should then lead to the question of whether an act of intelligence/awareness can or does produce it. Of course, we are unable to demonstrate scientific evidence of this at this time. Nevertheless, psychology has increasingly become more comfortable with the existence of the unconscious and the subpersonalities that act upon the person. Of course, it is quite a jump to a consideration of the unconscious aspects of a person communicating via the reversed sounds of the speaker. Perhaps in time, further understanding of the mysteries of the brain and the unconscious will lead toward more consideration of the phenomenon.

Byrne, T & Normand, M. (2000). The demon-haunted sentence: A skeptical analysis of Reverse Speech. Skeptical Inquirer (24)2. Retrieved from haunted_sentence_a_skeptical_analysis_of_reverse_speech1

Analysing speech reversals can be interesting as Reverse Speech is still a mystery. I have amused myself by doing this with a Trump reversal which is on YouTube

Documented RS: I’ll see in the lawsuit    

FS: do solemnly swear

Breakdown of phonemes and syllables

lawsuit – [do sol]emnly   Language clearly follows FS phonemes

see in the – sol[emnly s]wear

  • Perception of ‘see’ comes from [y s].
  • Perception of ‘in the’ has come from [emnl]. The two syllables come from the movement from [em] to articulation of /l/. Of course, /n/ is silent in ‘solemnly’.
  • It is possible to perceive [n] as articulation moves through /l/ to onset of /m/ in the RS. The articulation of [l] on the alveolar ridge helps give perception of a short ‘in’, although /l/ can still be heard as well.
  • A close listening reveals the sounds as ‘seal mo’. The /l/ is quite clear, and ‘the’ is ‘mo’. As /m/ may be perceived as [n] sometimes, this perhaps has helped give the perception of ‘the’, Of course even in FS, ‘the’ is not always articulated clearly and can sound like different things on close examination.  
  • So, primarily the issues to consider in this section is the existence of an /l/ and the ‘mo’. In an overall listening it can certainly sound like the documented reversal; this is helped by the low energy of /m/in ‘mo’, as well as its prosody and duration of perceived syllables (apart from documented recorded ‘I’ll’).

I’ll – s[wear]

Consideration of ‘you’

Now let’s take for argument sake that it is meant to be ‘see in the’. One may wonder if there is a short ‘you’ also, to make ‘I’ll see you in the lawsuit’. One can ‘project’ its existence into it. In natural speaking small words can be shortened and spoken very quickly and be missed. I’m not claiming it is there. I’m not trying to prove that it is.

Let’s assume ‘the’ is meant to be there …. To consider ‘you’, we need to focus on the ending of ‘see’ where there is some issue. An /l/ can be heard whether it is isolated with ‘see’ (‘seal’) or isolated slightly later (‘iln’). But, as the vowel moves toward /l/, is this meant to represent ‘you’? If we isolate ‘seal’ we hear the diphthong vowel with the short /i/ first, and a rounding from the ‘al’. Isolating the section from slightly before the rounding, and capturing the end of /i/, we can hear something quite similar to ‘you’. Yet, it is still easy to perceive the /l/. We can isolate sounds in different places and hear different words, which are not real, so we need to be careful. After this, /n/ can still be perceived in the release of /l/ and onset of /m/.

So, can ‘you’ be there? Or is /l/ meant to represent ‘in’? Or, is it ‘seal’? We can miss something subtle entirely, or we will put it in there because it makes sense to the grammar in our brain. What we need to do is look for indications that it is, in fact, there.

So, in the wash-up I certainly couldn’t say it is there. Like in FS, not all sections of language come out clearly and can sound like different things on close examination. I believe in examining for all possibilities.

Could the reversal say something else?

Within in this reversal one may hear ‘seal my loss’/’see in my loss’/’see in the loss’.

We need to look at where the reversal begins and ends. As ‘I’ll’ does not naturally sound like the beginning of the reversal as it has been documented in the recording (different time duration relative to the rest, for example), we either begin it at ‘see/seal’ or assume it is ‘Her, I w’ll’ or ‘hurl’.

If we assume the documented ‘see in the’ without ‘you’ is correct, then we have ‘Her, I w’ll see in the loss/Hurl. See in the loss’. However, phonemically it is also, ‘Seal my loss’, but may be ‘See in my loss’. The rounded vowel after [m] is not really an issue, and could be ‘my’. However, we need to see where the ending of the reversal lies. The last part, ‘suit’, does occur naturally as part of this reversal, so it does seem to end at ‘lawsuit’. So, we get something like ‘Seal my lawsuit’.

If ‘Her, I w’ll’,  then ‘her’ is identifying a woman, then saying what he will do. If it were ‘hurl’, it may represent disgust, or fear and anxiety (see commentary about other reversals below).

A general listening does give a perception of ‘in the lawsuit’. And that is how we interpret what people say as we listen to them speak – overall, big picture listening.  We would have to assume the alveolar articulation of /l/ is meant to be /n/ (and there is some perception of /n/ in the release of /l/ and onset of /m/), and the /m/ is meant to be /n/ to produce ‘the’. Of course, in natural speaking, when the previous word ends in /n/, the /th/ quite naturally takes on /n/ characteristics.

But because of the two issues, we should consider alternatives as well.

Her, I w’ll see in my lawsuit/Hurl. See in my lawsuit/Her, I w’ll seal my lawsuit/Hurl. Seal my lawsuit/  

One more thing I will add here. The lawsuit has been applied to the one publicly occurring at the time of writing this – the overturning of Trump’s migrant ban and Trump’s legal counter. However, this occurred at the time of inauguration, and I understand there had been other lawsuits occurring from Trump’s business activities, then. One does not have to see it as predictive and applying to the current issue. This is like people interpreting the words of some prophet or psychic from past history as pertaining to occurrences in their generation. But of course, at the same time, it is quite easy to see Trump entertaining future lawsuits from his actions that would come. Perhaps he was thinking about Hillary! It just doesn’t mean it is a psychic occurrence.

What about other reversals with it, and other voices?

Because it’s great fun investigating the possibilities of RS, let’s take a look at language-sounding pieces aside the reversal.

Before the reversal (in the RS direction, after in the FS direction), we may hear something like:

Ill of fear, flow our/out (Dad)?  

[that I will faithfully]

Ill – faithfu[lly] – closest to ‘ill’; lack of /h/ for ‘heal’. ‘Ill’ more appropriate anyway.

of – faith[fu]lly

fear – f[aith]fully  [th] can be perceived, but [f] can also be perceived as an alternative.

Flow out – tha[t I will f]aithfully   Sounds follow the FS phonemes clearly.   

What may be heard as ‘Dad’ is not natural with the previous words in stress/amplification/pitch etc, and it could be counted as gibberish. If not, perhaps Trump’s unconscious added his father to the mix.

So we have something like:

Ill of fear. Flow out.

Before the above example in the RS direction (last in the FS direction) is:

Weak, you’re sca ….   

faithfully [execu]te

weak – exe[cu]te   rounded vowel and [c] produce ‘weak’.

you’re sca … [exe]cute  [ex] produces ‘sca …’ and [e] produces ‘you’re’.  

You will notice the great difference in voice type between ‘Ill of fear. Flow out’ and ‘Seal my lawsuit’/’Seal in the lawsuit’. ‘Seal my lawsuit’/’See in my lawsuit’ is Trump’s normal voice, while the other one is very different. Because it is Trump’s normal voice, we could assume it is an aspect of Trump that is part of his ego self, that is, part of the personality that creates Trump’s outward identity. ‘

I have found that one may hear a series of reversals that flow directly on from one to another, like a conversation occurring between selves. One ‘voice’ can quickly take over from another without pause between. In this case, the first voice did not get to complete what it was saying – which possibly was ‘Weak, you’re sca[red]’. There is a sibilant-like sound occurring at the beginning from the [t] in ‘execute’, however, this can be seen as an extraneous sound, and therefore the reversal begins at ‘weak’. The second voice, took over with ‘Ill of fear; flow out’.

Should we follow the FS direction or the RS direction for sequence of reversals? If the interrupted reversal is genuine, then it must have come first at least in relationship to the one immediately following I the RS direction. The first one to occur in the FS direction – ‘seal etc’ may have come first in the RS, especially if ‘Dad’ and/or ‘hurl’ is gibberish. If these are not gibberish, then one may consider a sequence from start to finish in the RS direction. This brings up the question of how long can reversals occur continuously in the RS direction? Probably several seconds anyway as this is how long ahead our brain can be formulating what is going to be said ahead of speaking.

There is a relationship between the two reversals – weak, scared, ill of fear. These then, would be different aspects of Trump , fearful of the situation he is in. Does ‘ill of fear’ mean that he feels ill because of fear, or that he is sick and tired of fear? If the first ‘voice’ was telling him that he is weak and scared, the next ‘voice’ may have hit back, interrupting, saying that he is sick and tired of fear and he is going to let it ‘flow out’, whether he means the fear flow out and leave him, or that he will exude confidence and lack of fear to the public. That is one interpretation.

In the wash up, listening to the whole utterance, the brain easily follows the articulated /l/ as ‘in’ because /l/ at the end of ‘seal’ is a hard /l/ and there is some perception of /n/.

Weak, you’re sca …/Ill of fear, flow out (Dad)/(Her I w’ll/Hurl) see in my lawsuit/see in the lawsuit/seal my lawsuit    

Recently I have been examining speech reversals on the internet. It is clear that there is a plethora of examples that are ambiguous or obviously incorrect. Clearly, projection is occurring. Yet, some responders on the various sites indicate agreement with examples while others oppose. I have to assume that concurrence with highly dubious reversals is a case of priming which Mark Newbrooke claimed was a factor in hearing reversals. This is where the words are given to the listener beforehand, and this influences the listener’s perception. The listener perceives a syllable count, hears some phonemes or allophonic variations that are within linguistic limits, and the remainder is projected into; the mind adds the necessary sounds, and/or deletes others so that a recognisable string of language occurs. Furthermore, there are many reversals that are of 2 to 5 words, Out of the multitudinous strings of speech occurring each day, it is not unreasonable to expect short language-like strings from the reversed sounds that are simply coincidental occurrences. 

So, projection occurs; priming occurs, too. Short strings of speech are not all that remarkable – often they are just the reverse of the sounds of the forward speech, and probably occur when different people say it, or there are a couple of ambiguous sounds that allows the mind to fill in the ‘right’ sounds. But, with all that said, does this mean that there is nothing genuine about reversed speech? Not necessarily. It means projection is an issue. It means that there are plenty of poor or ‘shaky’ examples around. 

Amongst all the valueless apparent language, there is some pretty powerful stuff, and stuff that occurs in much longer strings of language, that I believe, are reasonably there within linguistic bounds. It is these longer strings that ‘separate the men from the boys’ so to speak. 

This is why I continue to look at Reverse Speech.

This study, published in 1993, chiefly looked at the effect of reversals on listeners, and is motivated by previous claims of influence of backward messages in rock music. It does not address Reverse Speech.

In the study, the authors played backward messages to one group and the forward speech content to another group. Afterwards, to both groups, the backward speech examples were played along with other examples which were not in the original examples. Their findings showed that the group that listened to the backward speech were able to recognise more of the actual backward speech played again to them than the ones that heard only the forward speech. As expected in the study, those who heard backward messages were able to recognise more than those who heard the forward speech of the backward examples played again. Acoustics (backward message group) were recognised more than meaning (forward speech group).

A second part to the study had groups read forward speech examples and decide whether they were what they had heard. The forward speech group recognised more of the speech than the backward speech group.This, of course, is understandable. Asking people to recognise the forward speech behind backward speech examples of a set of examples they heard once is asking their conscious mind to recognise something supposedly subconscious, soon after hearing. The conscious mind gets in the way, of course. The subconscious is sub … conscious. Play forward speech statements and then ask a group to recognise typed statements they had heard … well, how significant will the result be compared to the other group? Tests of this sort are only really worthwhile through a longer term observation of behavioural change through the influence of the subconscious. Yet, even then, those who listened to the backward speech recognised the forward statements at a proportion of 0.53 compared to the forward speech listeners at a proportion of 0.55, which is not a great difference.

In any consideration of Reverse Speech, there is little that is particularly informative here – the effect of reversals on the non-speaker listener is of interest, of course; however it is secondary to the consideration of whether real speech occurs, and what this means for the speaker himself. Secondly, asking the conscious mind to simply recognise what the subconscious is supposed to recognise, is just as likely to produce error, and, therefore, having those who heard forward speech and then seeing the typed forward speech, will, naturally, have an outcome greater than one who did not hear it. Thirdly, in Reverse Speech, intelligible language occurs in both modes. This study refers to forward speech played in reverse as gibberish.

A further point – the authors note that subjects were inexperienced at listening to backward speech. They suggested that people who were skilled at converting forward speech into backward speech (and vice versa?) could be good candidates for detection of backward messages. Of course, consciously-aware conversions have little to do with subconscious influences of messages. Again, there is nothing in this that has value in any consideration of Reverse Speech.

Prompted language

The authors state, “the target may be missed many times unless listeners are warned to listen for it”. This is an argument brought up by other researchers such as Newbrook and Curtain (1997) – that suggestion prompts hearing. Their point is quite correct –  suggestion can and does manipulate what people hear/believe etc. And, in regard to listening to backward messages, there is validity in the claim. However, it becomes a case of this can happen, therefore in all situations and cases it can be nothing more than suggestion.

The authors put forward the premise that potential meaning is there, but actualised meaning is not unless it can be heard by the listener. So, for meaning to be actualised, it must be decoded by a listener; therefore, if a listener fails to decode it, then there is no meaning. On the other hand, meaning is simply potential unless there is a listener with the right skill/ability can decode it, such as in understanding a foreign language. Their implication seems to be that, unless listeners can pick up messages independently, free from prompting, the message remains potential, and not actualised; in other words it may not be there at all. A message, then, may not be heard as something easily comprehensible,  and it would require either a particular decoding skill, or someone pointing out that it is there, aka suggestion.

Surely, if a backward message that approximates language is played, it stands to reason that more people will hear the message if pointed out than if not indicated. A study by Thorne and Himelstein (1984) found that 18% of listeners who were told to listen for messages in rock music (but not specifically satanic messages) heard satanic words, while 41% who were told to listen for satanic messages heard them. Of course, people will not always pick up short snippets of language that appear suddenly and unannounced, and that are not well enunciated – if two or three words from rock music, or from fast, casual speech were played amongst gibberish, surely more will pick up language that has been first indicated?

So, yes, there is no argument that people are influenced by verbal suggestion. Yet, if one played short segments of backward messages amongst gibberish, it is understandable that people will be more likely to recognise it if prompted. This, quite frankly, may be found to be true for forward speech. Play rock music or fast casual speech as short snippets of three to six words, and place it amongst gibberish, which group will hear significantly more? The prompted or unprompted group? The fact that more will hear speech that is prompted than unprompted does not ‘prove’ the lack of existence of backward messages. In addition, it must be realised that backward speech is not exactly the same as forward speech. Although many linguistic processes are shared by both modes, backward speech has its own nuances that occur through the subconscious communicating at the same time and space as the conscious speaker. People are not used to hearing speech in its backward mode. Just as people do not hear some accents as well as others, people will be less likely to pick up speech easily (and unprompted) that can be different to what they are used to hearing.

There are three questions that inform my thinking about Reverse Speech.

The first question is, are there substantial amounts of grammatically acceptable, linguistically-viable language in the reverse of speech? The answer is yes, there is. This can be proven. The next question is, can this language inform us about the person? The answer is yes, it does appear so. There is substantial evidence that is contextually/person-relevant (although I admit this can also be a product of projection and wrongful interpretation). The next question is, who/what is producing this language, what is the situation actually referred to, and, who does the message refer to? Any worthwhile interpretation would, and needs depend upon which subconscious aspect is communicating (if one is to differentiate aspects rather than see the subconscious as a single entity), and the actual situation/experience referred to, and who the referent is.

Begg, I. M., Needham, D. R., & Bookbinder, M. (1993). Do backward message unconsciously affect listeners? No. Canadian Journal of Experimental Psychology, 47(1).

Newbrook, M., & Curtain, J. (1997 ). Oates’ theory of Reverse Speech: a critical examination. The Skeptic (17)3

Thorne, S.B., & Himelstein, P. (1984). The role of suggestion in the perception of satanic messages in rock-and-roll recordings. Journal of Psychology (116), 245-248