Code-Switching (Style-Shifting): When Japanese Speakers Mix Registers
In this article, code-switching in Japanese means a speech style shift between です・ます and plain form that one speaker makes inside a single conversation. Often, that shift does not reset the relationship that set the baseline.12 For a late-intermediate learner who already controls both registers, the next skill is reading the movement between them: what a sudden plain predicate or a sudden です・ます is signaling.
Overview
Every Japanese conversation has a baseline speech level. A senior gets です・ます; a close friend gets plain. Style-shifting happens when a speaker steps off that baseline for a moment, then usually steps back.32
Japanese sociolinguistics calls this speech style shift, スピーチレベルシフト (speech-level shift), or スタイルシフト (style shift): one speaker temporarily switching between speech levels (丁寧体／普通体) within one conversation.23 This article focuses on decoding the shift, not forming either register.
The component forms (です・ます versus plain/だ) are N5–N4 grammar. Reading what a mid-discourse shift between them signals assumes prior control of both registers and of keigo, which is late-N3 to N2 territory.4 There is no single JLPT can-do statement for style shift; it is a discourse skill rather than a list item.
What "code-switching" means here (and what it does not)
The phenomenon here is intra-language speech style shift: one speaker alternates between です・ます (polite, 丁寧体) and plain or だ (普通体) forms inside one conversation.12 Liu (2013), following earlier work, defines スピーチレベルシフト as switching between different speech levels within a discourse; Maynard (1991) frames the same data as the mixture of da and desu/masu forms within a single turn.21
This is not bilingual code-switching. In its established sociolinguistic sense, code-switching is the alternation between two languages. Poplack (1980) defines it as "the alternation of two languages within a single discourse, sentence, or constituent."5 Myers-Scotton (1993) defines it as a bilingual's selection of forms from an embedded variety inside utterances of a matrix variety.6
The です・ます↔plain phenomenon is intra-language register variation. Japanese literature labels it speech style shift, not code-switching (コードスイッチング).21 This article keeps "code-switching" as the search-facing term but means style-shifting throughout.
Maynard's data span casual conversation, fiction dialogue, and literary essay, and the mixture appears in all three, so it is not an artifact of one genre.78
Baseline register vs. a momentary shift
The core model is simple. A speaker holds a baseline speech level for a given relationship, such as です・ます with a senior. A shift is a momentary deviation from that baseline, not a renegotiation of it.32
Ikuta (1983) treats the shift as a conversational strategy: the speaker moves away from the baseline to do interactional work, then returns.3 Liu (2013) operationalizes this with the term 基調 (base level) and tracks each shift as a departure from and return to it.2
The loop below is the shape to hold in mind: the baseline persists, the shift is the bump, and the return is the default.
The baseline can itself be plain. Among casual friends (友人同士), the default is plain with momentary upshifts to です・ます. Toward a senior, the default is polite with momentary downshifts to plain. Liu's data show both directions inside the same three-party conversation.2
How this differs from deciding to switch registers
A style shift is not the deliberate, relationship-level decision to drop into casual speech, such as agreeing to use タメ口. Ikuta (1983) characterizes the shift as an attitudinal, moment-to-moment movement of empathy or distance, not a resetting of the social relationship.3
Maynard frames the plain insertion as reflecting the speaker's internal perspective and involvement at that instant, not a change in how the speaker ranks the addressee.18 The deliberate, relationship-level decision is a separate skill with its own treatment.
The mechanics of a shift
Downshift: です・ます → plain
In a polite-baseline conversation, the common direction is the downshift: an utterance, or just its predicate or final element, drops from です・ます to plain/だ.12 Liu (2013) found these downshifts clustering in two content types: one-sided expression of the speaker's own opinion or feeling, and repetition or echoing of another's words (繰り返し).2
Maynard reads the da style as marking the speaker's internal perspective and involvement, yielding an informal, close, casual tone.18
In the example below, the baseline is です・ます because the speaker is talking up to seniors. The plain tail …ないよ is the downshift point.
何でもかんでも女子をつけりゃいいわけじゃないよ。2
"It's not like slapping '-girl' onto absolutely everything works, you know." (baseline です・ます, downshifted to plain on the final predicate)
Here the speaker asserts his own opinion about a third party's word choice. Liu reads the plain tail as low orientation to the listener, or 聞き手を特に意識していない (not particularly conscious of the listener).2 The second downshift type, repetition, works the same way on the surface.
{笑い}そこがラジオ。2
"(laughs) That's the radio part right there." (echo of the prior speaker, dropped to plain)
In Liu's three-party data the downshift is typically a single utterance or a single predicate inside an otherwise です・ます turn; the whole turn does not flip.2
Upshift: plain → です・ます
The marked, less frequent direction is the upshift: from a plain baseline, or after a run of plain forms, back up to です・ます.23 Liu (2013) found these upshifts in three contexts: set phrases that signal the end of a conversation, criticism of the addressee, and presentation of an opposing stance.2
Her attributed function is that the speaker deliberately places psychological distance (心的距離をわざと置く) and so holds firm to his own position.2 This matches Ikuta's [+Distant] pole, where です・ます indexes attitudinal and social distance. Moving up to it therefore re-marks distance.3
In the next example, the speaker has just been laughed at by the others. He upshifts to a polite request, putting distance between himself and them.
ちょっと聞いてください。2
"Hold on, would you listen to me a second." (plain baseline upshifted to the polite …ください)
Liu labels that one criticism of the addressee and marking of distance.2 The opposing-stance type looks like this:
2対1ですよ、今。2
"It's two against one right now, you know." (upshift to …ですよ inside a contested exchange)
Maynard's account predicts the upshift symmetrically: です・ます is the high-awareness-of-addressee form, so moving up to it foregrounds addressee-awareness and distance.18
Where shifts surface in the sentence
Shifts surface at the sentence-final predicate and on final-particle-bearing tails. They do not usually flip a whole turn. Liu classifies forms by their tails: です・ます tails (です／ます／でしょう／てください) and です・ます-plus-particle forms (ですね／ですよ), versus だ tails and だ-plus-particle forms (いいね／あるよ).2
Embedded and quoted clauses carry plain form independent of the matrix register. A quoted clause stays plain even when its framing is polite. Maynard notes that inserting quotation-like material lends a vivid, directly quoted, foregrounding effect.78
Because shifts land on the predicate tail, the gendered and stance indexing carried by final particles (よ／ね／わ／ぞ／ぜ) often moves with the shift. SturtzSreetharan (2006) shows Japanese men using sentence-final particles and address terms to construct stance, seniority, and masculinity. That is the layer the shift's tail sits on.9
What a shift signals
This section is the decoding guide. Each function below is tied to a named theorist. The verbatim examples come from Liu's transcripts, and the function each illustrates is attributed to the primary source for that function.
The map below splits the two directions into their dominant signals. Read it as a lookup: first hear the direction, then narrow by content.
Emphasis and emotional involvement
A downshift to plain marks heightened feeling, involvement, or a punchy, vivid assertion. Maynard reads the da form as expressing the speaker's internal perspective and emotional involvement in the narrative.18 In Liu's earthquake retelling, the speaker drops to plain across a blow-by-blow re-enactment, pulling the listener into the scene.2
すごい。2
"It was unreal." (bare plain exclamation inside an otherwise です・ます retelling)
Inside that polite-baseline turn, the bare plain exclamation marks peak involvement, not a relationship change.2 The function (involvement and affect) is Maynard's.1
Self-directed asides and 独り言
Self-directed, soliloquy-like speech defaults to plain even inside a polite conversation, because the speaker is not orienting to an addressee. Hasegawa (2003, 2010) shows that in soliloquy there is no addressee. As a result, addressee-honorifics (です・ます) and addressee-oriented particles do not appear; the plain form is used.1011
Maynard frames the same effect as a da insertion reflecting the speaker's internal, thought-prevailing perspective with low awareness of the addressee.18 Liu's downshift function 聞き手を特に意識していない (not particularly conscious of the listener) is the conversational counterpart.2
This is why a quiet plain mutter inside a polite conversation, such as あれ、変だな ("Huh, that's strange"), is heard as thinking aloud rather than rudeness: it is momentarily addressee-less speech.1011
Intimacy, solidarity, and rapport
A brief downshift can be a bid for closeness or solidarity. Ikuta (1983) finds that a shift toward the [−Distant] (plain) level coincides with attitudinal closeness and a flow of empathy; speakers shift from [+Distant] to [−Distant] when showing strong agreement or positively evaluating what was just said.3 Maynard ties the da style to feelings of closeness and in-group belonging.18
いいね。2
"Oh nice." (plain agreement tail of the empathy/closeness type)
This is a plain positive-evaluation tail of the kind Ikuta links to empathy. Liu lists だ-plus-particle forms like いいね among the plain-style tails in friend talk.2 The function (empathy and closeness) is Ikuta's.3
Distance, reframing, and "back to business"
An upshift re-marks distance, ends an aside, or signals a frame or role change. Ikuta treats [+Distant] (です・ます) as the speaker's attitudinal distance, and treats shifts as signals of structural transitions in the talk.3 Liu's だ → です・ます upshift deliberately places psychological distance and marks boundaries such as set-phrase closings (会話終了時の合図, signals that a conversation is ending).2
Geyer (2008) adds the institutional layer: in workplace and faculty meetings, です・ます and standard forms index "on-stage," depersonalized, audience-directed speech. Plain and dialect forms index "off-stage," personalized speech.1213 The upshift is the move back on-stage.
はい、以上です。2
"Right, that's everything." (set-phrase です・ます upshift used as a closing signal)
This is the wrap-up frame move, a set-phrase upshift used as a closing signal.2 The function (boundary and distance) is Ikuta's and Geyer's.312
Humor, quotation, and reported speech
Plain-form quotation or repetition inside polite narration creates a vivid, dramatized, sometimes comedic effect. Liu's repetition (繰り返し) downshift functions like quoting the other person's words.2 Maynard notes that optional quotation-like insertion makes the proposition resemble directly quoted discourse, adding a vivid effect of saying.78
A directly quoted clause stays plain regardless of the polite frame around it. Reported speech therefore routinely produces a local register dip that is not a stance change toward the addressee.7 This is the mechanism behind register play in storytelling and broadcast talk: the narrator stays polite to the audience while voicing quoted or internal lines in plain form for liveliness.72
Nuance and usage contexts
One turn, two levels
A single stretch of talk can open polite and close plain, or the reverse. Liu documents です・ます-baseline turns that contain だ predicates, and plain-baseline turns that contain です・ます, within the same speakers and the same conversation.2
In her earthquake retelling, one of the speaker's turns opens in です・ます and slides into plain as he gets absorbed in re-enacting the scene.
そうです。… テレビはひっくり返るわー。2
"That's right. … and the TV's going right over." (opens です・ます, slides to plain as involvement rises)
The shift here signals involvement, not a relationship reset. The lines between are omitted in this two-line version; the full transcript is in Liu's 会話 7.2
Conscious vs. unconscious shifts
Most shifts are automatic, not errors. Maynard's central argument is that the mixture is systematically motivated by discourse modality, meaning perspective, discourse organization, and interaction. It is not sloppy speech.17
Some shifts are strategic. Ikuta and Liu both describe the shift as a deliberate conversational strategy, such as upshifting to hold one's ground.32 Either way, "they made a mistake" is the wrong interpretation.
Cook (1998, 2008) argues that the masu form indexes both addressee-deference and the speaker's self-presentation. Because of that, a shift out of masu marks a stance change: the speaker steps out of the self-presentational, on-record footing.1415 That stance change can be monitored or automatic, but it is meaningful either way, not an accident to be corrected.
Reading shifts in real input
In TV, radio, podcasts, anime, and natural conversation, the practical skill is hearing the baseline and catching the deviation. Style-shift studies are built on these kinds of speech: broadcast talk, classroom talk, workplace meetings, and casual multi-party conversation.12132
Listen for a sudden plain predicate or plain final particle inside polite talk. That downshift may signal involvement, self-talk, quotation, or rapport. Also listen for a sudden です・ます inside casual talk. That upshift may signal distance, criticism, or wrap-up.2312
Good to know
Do not imitate every shift you hear
Comprehension comes before production. A downshift that a native speaker uses automatically, for involvement or self-talk, can read as rude or over-familiar when a learner inserts it at the wrong moment. The reason is that the masu form being dropped carries self-presentation and deference value.1415 Cook (2008) documents learners being explicitly and implicitly socialized into when the shift is licensed.15
Producing native-like shifts is a late skill. The shift's meaning lives in the deviation from the baseline. That assumes the learner can first hold a stable baseline reliably.153
A shift is not a register reset
A common misreading treats a momentary shift as a permanent register change: "He used plain form once, so the relationship is casual now and I can use タメ口." The baseline です・ます relationship still holds. The single plain form was a momentary involvement, self-talk, or quotation marker that does not reset anything.132
Ikuta and Maynard both model the shift as a deviation from a held baseline that returns to that baseline. Liu likewise tracks shifts as departures from a 基調 (base level), never as a new baseline.312
Gendered and final-particle layers
Shifts ride on sentence-final particles, which carry their own gendered and stance indexing. Because the shift lands on the predicate tail, the attached particle (よ／ね／わ／ぞ／ぜ／の) layers gendered and stance meaning on top of the level shift.29
SturtzSreetharan (2006) shows men using sentence-final particles and address terms to construct seniority, masculinity, and stance in real conversation. A style shift surfaces in the same tail position, so the two indexing systems co-occur.9 The gendered-language layer is worth studying alongside this one.
"Down for the heart, up for the wall"
A retention phrase for the two directions: down for the heart, up for the wall. Down to plain means the speaker's heart, involvement, or inner thought shows through (Maynard's internal perspective, Hasegawa's addressee-less self-talk).110 Up to です・ます means the speaker raises a wall of distance (Ikuta's [+Distant], Liu's 心的距離をわざと置く, deliberately placing psychological distance).32
The phrasing is a mnemonic; the underlying functions are the sourced ones, mapped onto their two dominant decodes.
See also
- Spoken-Word vs. Written-Word Japanese: 話し言葉 vs. 書き言葉
- Formal Written Japanese (である調): The Register
- Uchi vs. Soto (内・外): The In-Group / Out-Group Axis
- Tatemae and Honne: Public Stance vs. Private Opinion in Japanese