A scoundrel who thinks evil of a slip of the tongue. Slips of the tongue are often amusing to listeners because they often leave much room for interpretation. But contrary to popular belief, they usually do not reveal the speaker’s secret thoughts. Rather, verbal gaffes provide a unique insight into how our speech production works. The exact process (from thought to utterance) is anything but easily accessible. Most of it escapes our awareness and takes place in fractions of a second. Slips of the tongue provide a window through which we can view the various stages of the speech process. As the saying goes, you never know how something works until it goes wrong.
When we want to send a message, our internal speech production system goes through a series of actions. We have to find the appropriate concepts and expressions and put it all into an orderly grammatical form, since we mostly speak in sentences. Only then is the plan put into a phonetic sequence. By and large, the speech planning system is highly error resistant, i.e. slips of the tongue are rare. A speech error occurs only once every 1,000 words, at an average rate of three words per second.
The most important findings about the architecture of our speech production are based on data from research on slips of the tongue. The basic assumption is that slips of the tongue do not occur at random, but follow certain rules. For example, we do not mispronounce at random; rather, linguistic mishaps always involve specific elements such as sentences, words, morphemes (the smallest meaningful word unit), syllables, or sounds. Moreover, it is almost never the case that we mix up function words (such as articles) and nouns when we get confused. Therefore, different levels of language planning are responsible for this. In order for two words to influence each other, they have to be on the workbench at the same time. On this basis, the individual stages of language planning and the organization of the mental lexicon can be deciphered step by step.
The latter is stored in long-term memory and contains 50,000 to 100,000 entries that we access in a fraction of a second when we speak. The larger the vocabulary, the greater the risk of making a mistake. For example, German is rich in idioms that have no literal meaning. If you have a lot of such idioms, you can easily get confused. When it comes to slips of the tongue, we can see how to look them up in this dictionary. The access is done in two separate stages, according to the meaning of a word and according to its phonetic form. The fact that meaning and form can fall apart is vividly demonstrated by the phenomenon of lying on the tongue: The content of what is to be communicated is known, but the phonetic form is not, or only fragmentarily.
Further evidence of this dual path is provided by two types of promises: contaminations and substitutions. In contaminations, which are common, words or phrases with similar meanings compete with each other and are often combined as a compromise. They tell us something about the organization of the internal lexicon: apparently, terms and concepts with similar meanings are stored close to each other and activated simultaneously. Otherwise, the linguistic mixtures could hardly arise. In the lexicon of forms, on the other hand, concepts that are similar in form, i.e., have comparable sounds, a comparable number of syllables, or endings, are closely related. Therefore, it can happen that the similar form is erroneously activated. In such a form-related substitution, “jetting around the world” quickly becomes “chatting around the world”.
How does our speech production proceed in time? And can the different stages (from the intention to express something to the audible articulation) influence each other? Slips of the tongue have also provided important clues. For example, in the sentence “Den musss man mit rohen Handgluhen anfassen” – a cross between “handle with kid gloves” and “treat like a raw egg” – something remarkable happens: Apparently, samt is considered an adjective, and thus a specification of glove, so that the word roh can take its place and is properly in the accusative plural.
It can be concluded that the error occurs when the sentence structure is not yet fixed and grammatical adaptation is possible. This early stage of language production is called the functional level. Here, the content information appropriate to the message is selected from the lexicon of meaning. The order of the words is not yet determined, and they do not yet contain sound information. However, the lexical units are already marked with clues about their part of speech and position in the sentence. At this stage, not only contamination but also content substitution occurs.
Even as we utter a first, we are looking for the next items in our internal plan and storing them in working memory. Normally, these items wait their turn. Occasionally, however, they change places or are recognized too early or too late. When swapping, individual words change their place (e.g., “He has a horse like a bit”). In a subform, the stranded-goods error, two word stems swap places, leaving the endings behind (e.g., “You are a pious lamb” instead of “You are a pious lamb”). But only superficially do all these swaps follow the same linguistic pattern.
Unlike flotsam and jetsam and sound swaps, word swaps only ever involve elements of the same part of speech changing places, such as nouns and nouns. This results in an incorrect assignment of subject and object. An example: “My cat has a sister”. In this case, the elements can be relatively far apart, as in the following example: “Men can still drink when they’ve driven something.” This means that there must be a stage in language planning where grammatical roles are assigned, but the final position of the words in the sentence frame has not yet been determined. Thus, word substitutions occur at about the same time as contamination and content substitutions – namely, at the functional level.
The next step is to formulate the sentences. After the sentence structure, the content words are determined and in the next step the function words are inserted. Then the phonetic information from the morphological lexicon is added. It is at this stage that the phonetic exchanges take place. Therefore, the elements in question are always close to each other. The stranded errors probably occur just before the phonetic exchanges, because the direct neighborhood between the units is not necessary here.
We should also mention anticipation and perseveration, which are subject to similar limitations. In anticipation, a planned sound or semantic unit is realized too early. The elements involved must come from the same syllable position, and the words involved may belong to different word types. In perseverations, elements are re-realized even though they should have been erased from memory.
This positional level is followed by another level: that of the sounds. From here the concrete sound information reaches the articulatory apparatus. Linguistic errors at this level are, for example, loadings, in which, for example, a part of the planned utterance disappears completely. But they are extremely rare.
The stages of speech production are sequential, but there are also certain interactions. For example, “traipsing through the kitchen” becomes “traipsing through the carriage” through anticipation. Obviously, the lexicon was quickly consulted before the utterance, so that an existing word was created – a rather rare mechanism called lexical control. Often, however, slips of the tongue come from the internal language planning system. Occasionally, slips are actually influenced by external factors, and thoughts enter the language system.
Between 50 and 60 percent of slips are audibly corrected. This is done in two steps: detecting the error and restarting. Speakers are also listeners to their own utterances and can correct their mistakes if necessary. Occasionally, corrections are introduced by formulas such as “oh dear” or “uh”, by laughter, or by pauses. These signal an error to the listener and buy the speaker time to reformulate. The control is probably a conscious mechanism, but the restarted speech planning is involuntary. However, this sometimes makes you more confused.
There is nothing pathological about slips of the tongue, unlike, for example, slips of the tongue caused by brain damage after a stroke. But some aphasic disorders follow the same pattern as slips of the tongue in healthy people. In semantic paraphasia, people have trouble accessing the contents of the lexicon of meaning. As a result, they often confuse words such as chair and table or key and cup. In phonemic paraphasia, phonetic errors occur. Tongue twisters also differ from classic slips of the tongue in that they occur only during articulation. A slip of the tongue also requires complete knowledge of the language. If there are gaps in word memory, malapropisms can occur, especially with foreign words. In this case, the speaker uses a similar sounding word, but with the wrong meaning.
In the end, though, most slips of the tongue go unnoticed. Because the message gets across the same way. When we speak to people in the same language, they are so familiar with the language’s word formation rules that they can unconsciously correct any slip.